From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kara Subject: Re: [PATCH 0/3] BDI handling fixes Date: Fri, 17 Sep 2010 14:52:45 +0200 Message-ID: <20100917125245.GA3342@quack.suse.cz> References: <1284677051-28564-1-git-send-email-jack@suse.cz> <20100916234742.GA26030@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , Jens@suse.cz, "Axboe , linux-fsdevel@vger.kernel.org To: Christoph Hellwig Return-path: Received: from cantor2.suse.de ([195.135.220.15]:56335 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752377Ab0IQMxh (ORCPT ); Fri, 17 Sep 2010 08:53:37 -0400 Content-Disposition: inline In-Reply-To: <20100916234742.GA26030@infradead.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Thu 16-09-10 19:47:42, Christoph Hellwig wrote: > On Fri, Sep 17, 2010 at 12:44:08AM +0200, Jan Kara wrote: > > > > Hi, > > > > series of these three patches fixes the warning in __mark_inode_dirty() > > which happens when I do e.g. touch /dev/zero. The first two patches should > > be obvious enough and probably worth merging independently of the third > > patch. The third patch is upto a discussion whether we want to solve the > > problem that way or differently. Christoph, I know we spoke at LSF that > > inode_to_bdi() could be a per-sb method but the current version of the > > patch seems clean enough to me that we could maybe go even without the > > special callback? > > Feel free to go with the simpler one. But what I think really needs to > be changes is the no writeback flag - it's exactly the wrong way around. > > Instead just add a flag to allow writeback for the block device and > fs-specific bdi structures. Agreed. I just think that I'll first make the flags right and then just mechanically flip NO_WRITEBACK and WRITEBACK... Honza -- Jan Kara SUSE Labs, CR