From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from imap.thunk.org ([74.207.234.97]:56285 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750888AbaK1FhC (ORCPT ); Fri, 28 Nov 2014 00:37:02 -0500 Date: Fri, 28 Nov 2014 00:36:56 -0500 From: "Theodore Ts'o" To: Jan Kara Cc: Linux Filesystem Development List , Ext4 Developers List , Linux btrfs Developers List , XFS Developers Subject: Re: [PATCH-v4 2/7] vfs: add support for a lazytime mount option Message-ID: <20141128053656.GI14091@thunk.org> References: <1416997437-26092-1-git-send-email-tytso@mit.edu> <1416997437-26092-3-git-send-email-tytso@mit.edu> <20141127131421.GE30152@quack.suse.cz> <20141127230016.GH14091@thunk.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20141127230016.GH14091@thunk.org> Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Thu, Nov 27, 2014 at 06:00:16PM -0500, Theodore Ts'o wrote: > Well.... it's not quite enough. The problem is that for ext3 and > ext4, the actual work of writing the inode happens in dirty_inode(), > not in write_inode(). Which means we need to do something like this. > > I'm not entirely sure whether or not this is too ugly to live; > personally, I think my hack of handling this in update_time() might be > preferable.... .... and this doesn't work because it breaks ext3/ext4's transaction handling, since the writeback thread could be racing against some transactional update of the inode. So I don't see a way of making the queue_io / move_expired_inodes approach to the 24 hour time being tractable. So the alternatives that I can see at this point is either give up on the 24 hour timeout, or we fall back to handling this in update_time(). - Ted