From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ipmail03.adl2.internode.on.net ([150.101.137.141]:64064 "EHLO ipmail03.adl2.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750794AbeEKXyJ (ORCPT ); Fri, 11 May 2018 19:54:09 -0400 Date: Sat, 12 May 2018 09:54:06 +1000 From: Dave Chinner Subject: Re: [PATCH] xfs: test mount vs superblock shrinker races Message-ID: <20180511235406.GA23861@dastard> References: <20180511021152.10803-1-david@fromorbit.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Amir Goldstein Cc: fstests , linux-xfs On Fri, May 11, 2018 at 08:41:58AM +0300, Amir Goldstein wrote: > On Fri, May 11, 2018 at 5:11 AM, Dave Chinner > wrote: > > From: Dave Chinner > > > > Test case for superblock shrinkers running while the filesystem > > is being set up and/or torn down and tripping over inconsistent > > state. > > > > Signed-Off-By: Dave Chinner --- ... > > Isn't it dangerous for the time being? Your fix commit message > says you got reports of OOPS. If you have a kernel with the mount_delay sysfs option, you will have a kernel with the fix. If you don't have mount_delay, the test won't run. > Not sure what should be the authoritative condition to remove dangerous > point kernel release with a fix?? I'm not going to add "dangerous" to new regression tests that have a fix pending anymore because nobody is doing maintenance tasks like sweeping out tests that aren't dangerous to upstream. Tagging new regression tests as dangerous just creates unnecessary technical debt. Cheers, Dave. -- Dave Chinner david@fromorbit.com