From owner-freebsd-fs@FreeBSD.ORG Thu Sep 29 15:59:36 2011 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id CCA3B106566C; Thu, 29 Sep 2011 15:59:36 +0000 (UTC) (envelope-from mckusick@mckusick.com) Received: from chez.mckusick.com (chez.mckusick.com [70.36.157.235]) by mx1.freebsd.org (Postfix) with ESMTP id A8BBB8FC19; Thu, 29 Sep 2011 15:59:36 +0000 (UTC) Received: from chez.mckusick.com (localhost [127.0.0.1]) by chez.mckusick.com (8.14.3/8.14.3) with ESMTP id p8TFxc63084067; Thu, 29 Sep 2011 08:59:38 -0700 (PDT) (envelope-from mckusick@chez.mckusick.com) Message-Id: <201109291559.p8TFxc63084067@chez.mckusick.com> To: Attilio Rao In-reply-to: Date: Thu, 29 Sep 2011 08:59:38 -0700 From: Kirk McKusick X-Spam-Status: No, score=0.0 required=5.0 tests=MISSING_MID, UNPARSEABLE_RELAY autolearn=failed version=3.2.5 X-Spam-Checker-Version: SpamAssassin 3.2.5 (2008-06-10) on chez.mckusick.com Cc: Garrett Cooper , freebsd-fs@freebsd.org, Xin LI Subject: Re: Need to force sync(2) before umounting UFS1 filesystems? X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 29 Sep 2011 15:59:36 -0000 > Date: Thu, 29 Sep 2011 17:40:43 +0200 > Subject: Re: Need to force sync(2) before umounting UFS1 filesystems? > From: Attilio Rao > To: Kirk McKusick > Cc: Garrett Cooper , freebsd-fs@freebsd.org, > Xin LI > > 2011/9/29 Kirk McKusick : > > > Thanks for providing a bit more of the history on this codepath. > > > > Since 9-stable has now been branched, I believe that the best path > > forward is to check this change into head and let it sit there for > > several months so that we can get some experience with it. If it > > causes folks problems we can back it out. If it does not cause > > problems, then we can MFC it to 9-stable. > > > > Does this seem like a reasonable approach? > > In general yes, but I'd like to understand why unmount should fail so > much with SU... do we do extended period with vfs_busy()'ed > filesystem? > > I need more context here, likely I'd need to look into the PRs too > before to give an informative answer. > > Attilio I am definitely not in a rush on this, so by all means take some time to look it over. The EBUSY unmount has been in its current state for several years, so I am fine with taking a few weeks to sort out the correct solution. Indeed, I am glad that Garrett has volunteered to do some more serious testing. If this general approach is not correct, I can put a hook in for just UFS so that it can have its historic behavior. As you have noted, the SU code has a lot of activity that gets done under the protection of vfs_busy. So it may be the only filesystem for which draining the vfs_busy lock during unmount is needed. Will you be at the EuroBSD conference next week? If so we can discuss this there. > Date: Thu, 29 Sep 2011 08:38:59 -0700 > Subject: Re: Need to force sync(2) before umounting UFS1 filesystems? > From: Garrett Cooper > To: Kirk McKusick > Cc: Attilio Rao , freebsd-fs@freebsd.org, > Xin LI > > > Does this seem like a reasonable approach? > > I'll give it a quick run through first on some machines this weekend, > with NFS, UFS, and ZFS. It seems like this could negatively affect a > number of users, so I want to make sure that it passes a smoke test > before committing directly to HEAD. > > Thanks! > -Garrett Thanks for doing these tests to help us find out if there are landmines in this change. Kirk McKusick