From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [PATCH] 2.5.x write_super is not for syncing Date: Tue, 03 Dec 2002 14:13:15 -0800 Sender: linux-fsdevel-owner@vger.kernel.org Message-ID: <3DED2C7B.135E8F0D@digeo.com> References: <3DED0EE2.ECE22229@digeo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: "linux-fsdevel@vger.kernel.org" , Chris Mason , "Stephen C. Tweedie" Return-path: Received: from digeo-nav01.digeo.com (digeo-nav01.digeo.com [192.168.1.233]) by packet.digeo.com (8.9.3+Sun/8.9.3) with SMTP id OAA20483 for ; Tue, 3 Dec 2002 14:34:57 -0800 (PST) To: Bryan Henderson List-Id: linux-fsdevel.vger.kernel.org Bryan Henderson wrote: > > >The intent is that ->sync_fs() should only be > >called for sync, unmount, etc. > > Why unmount, etc.? That's just more assumption-making by FS as to what > these operations require for every filesystem type. > > unmount, which in general requires a rather different kind of syncing than > sync(), already has its own VFS call. > I don't know what etc. would be but it's probably also something a > filesystem driver might want to distinguish from sync(). Well, I'd say that telling the fs about umount activity is "out of scope" for this exercise ;) It seems reasonable to have a super_op which says "sync everything to disk". And to call that in response to /bin/sync and unmount. Doesn't it?