From: Christoph Hellwig <hch@lst.de>
Subject: Re: [PATCH v6 0/4] ext4: Coordinate data-only flush requests sent by fsync
Date: Tue, 30 Nov 2010 17:43:01 +0100
Message-ID: <20101130164301.GB30858@lst.de>
References: <20101129220536.12401.16581.stgit@elm3b57.beaverton.ibm.com> <20101130113906.176ffcad@notabene.brown>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: "Darrick J. Wong" <djwong@us.ibm.com>,
	Jens Axboe <axboe@kernel.dk>, "Theodore Ts'o" <tytso@mit.edu>,
	Andreas Dilger <adilger.kernel@dilger.ca>,
	Alasdair G Kergon <agk@redhat.com>, Jan Kara <jack@suse.cz>,
	Mike Snitzer <snitzer@redhat.com>,
	linux-kernel <linux-kernel@vger.kernel.org>,
	linux-raid@vger.kernel.org, Keith Mannthey <kmannth@us.ibm.com>,
	dm-devel@redhat.com, Mingming Cao <cmm@us.ibm.com>,
	Tejun Heo <tj@kernel.org>, linux-ext4@vger.kernel.org,
	Ric Wheeler <rwheeler@redhat.com>,
	Christoph Hellwig <hch@lst.de>, Josef Bacik <josef@redhat.com>
To: Neil Brown <neilb@suse.de>
Content-Disposition: inline
In-Reply-To: <20101130113906.176ffcad@notabene.brown>
Sender: linux-ext4-owner@vger.kernel.org

On Tue, Nov 30, 2010 at 11:39:06AM +1100, Neil Brown wrote:
> Would it not be sufficient to simply gather flushes while a flush is pending.
> i.e
>   - if no flush is pending, set the 'flush pending' flag, submit a flush,
>     then clear the flag.
>   - if a flush is pending, then wait for it to complete, and then submit a
>     single flush on behalf of all pending flushes.
>
> That way when flush is fast, you do a flush every time, and when it is slow
> you gather multiple flushes together.

We can even optimize the second flush away if no other I/O has
completed since the first flush has started.  That will always be the
case for SATA devices as the cache flush command is not queued.