From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932394Ab1EJNwE (ORCPT ); Tue, 10 May 2011 09:52:04 -0400 Received: from cantor.suse.de ([195.135.220.2]:34246 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753937Ab1EJNwC (ORCPT ); Tue, 10 May 2011 09:52:02 -0400 Date: Tue, 10 May 2011 15:52:00 +0200 From: Jan Kara To: Wu Fengguang Cc: Jan Kara , Andrew Morton , LKML , Mel Gorman , Dave Chinner , Christoph Hellwig , "linux-fsdevel@vger.kernel.org" Subject: Re: [PATCH 03/17] writeback: introduce writeback_control.inodes_cleaned Message-ID: <20110510135200.GH4402@quack.suse.cz> References: <20110506030821.523093711@intel.com> <20110506031612.167620443@intel.com> <20110506143614.GF18982@quack.suse.cz> <20110510022336.GB6758@localhost> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20110510022336.GB6758@localhost> User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue 10-05-11 10:23:36, Wu Fengguang wrote: > On Fri, May 06, 2011 at 10:36:14PM +0800, Jan Kara wrote: > > On Fri 06-05-11 11:08:24, Wu Fengguang wrote: > > > The flusher works on dirty inodes in batches, and may quit prematurely > > > if the batch of inodes happen to be metadata-only dirtied: in this case > > > wbc->nr_to_write won't be decreased at all, which stands for "no pages > > > written" but also mis-interpreted as "no progress". > > > > > > So introduce writeback_control.inodes_cleaned to count the inodes get > > > cleaned. A non-zero value means there are some progress on writeback, > > > in which case more writeback can be tried. > > > > > > about v1: The initial version was to count successful ->write_inode() > > > calls. However it leads to busy loops for sync() over NFS, because NFS > > > ridiculously returns 0 (success) while at the same time redirties the > > > inode. The NFS case can be trivially fixed, however there may be more > > > hidden bugs in other filesystems.. > > I think Christoph didn't like this patch > > (https://lkml.org/lkml/2011/5/4/123) and suggested that inodes_cleaned > > should remain local to fs-writeback.c... > > Yes, he didn't like introducing one more field to writeback_control, > which will be removed in patch 14. (It will be a lot of fuss to move > patch 14 here or move this logic after patch 14.) Yes, I've noticed (after I've sent this email) that you eventually remove the code later in the series :). So I wonder would it be that much work to just drop this patch? It's kind of dumb to add a new code which is removed a few patches later. But I'm stopping bitching about this now ;). Honza -- Jan Kara SUSE Labs, CR