From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kara Subject: Re: [PATCH 1/6] jbd2: Issue cache flush after checkpointing even with internal journal Date: Fri, 10 Feb 2012 14:58:12 +0100 Message-ID: <20120210135812.GJ10509@quack.suse.cz> References: <1326241872-20142-1-git-send-email-jack@suse.cz> <1326241872-20142-2-git-send-email-jack@suse.cz> <20120111124918.GD26337@quack.suse.cz> <20120209030551.GH18461@thunk.org> <4BF54606-BF0E-4CA8-98F7-25B16BCDA497@mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , linux-ext4@vger.kernel.org To: Theodore Tso Return-path: Received: from cantor2.suse.de ([195.135.220.15]:58442 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757955Ab2BJN6N (ORCPT ); Fri, 10 Feb 2012 08:58:13 -0500 Content-Disposition: inline In-Reply-To: <4BF54606-BF0E-4CA8-98F7-25B16BCDA497@mit.edu> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu 09-02-12 00:26:11, Ted Tso wrote: > One more thought. I wonder if we should be trying to call a version of > journal_cleanup_tail after we do a commit, since at that point we will > have done a flush so updating the journal superblock should be "cheap". Yes. That is in patch 6/6 ;). I just put it in a separate patch because the first patch is a correctness fix which was complex enough and this is just an optional optimization... Honza -- Jan Kara SUSE Labs, CR