From mboxrd@z Thu Jan  1 00:00:00 1970
From: Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [patch 6/6] mm: fsync livelock avoidance
Date: Thu, 11 Dec 2008 13:51:40 -0800
Message-ID: <20081211135140.89fac18b.akpm@linux-foundation.org>
References: <20081210072454.GB27096@wotan.suse.de>
	<20081210074209.GG27096@wotan.suse.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Cc: linux-fsdevel@vger.kernel.org, mpatocka@redhat.com
To: Nick Piggin <npiggin@suse.de>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from smtp1.linux-foundation.org ([140.211.169.13]:33814 "EHLO
	smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK)
	by vger.kernel.org with ESMTP id S1757440AbYLKVwA (ORCPT
	<rfc822;linux-fsdevel@vger.kernel.org>);
	Thu, 11 Dec 2008 16:52:00 -0500
In-Reply-To: <20081210074209.GG27096@wotan.suse.de>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Wed, 10 Dec 2008 08:42:09 +0100
Nick Piggin <npiggin@suse.de> wrote:

> This patch fixes fsync starvation problems in the presence of concurrent
> dirtiers.

One reject in xfs, which now does:

int
xfs_flush_pages(
	xfs_inode_t	*ip,
	xfs_off_t	first,
	xfs_off_t	last,
	uint64_t	flags,
	int		fiopt)
{
	struct address_space *mapping = VFS_I(ip)->i_mapping;
	int		ret = 0;
	int		ret2;

	if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
		xfs_iflags_clear(ip, XFS_ITRUNCATED);
		ret = filemap_fdatawrite(mapping);
		if (flags & XFS_B_ASYNC)
			return -ret;
		ret2 = filemap_fdatawait(mapping);
		if (!ret)
			ret = ret2;
	}
	return -ret;
}

ie: it returns a +ve integer on error, heaven knows why.

So I did

int
xfs_flush_pages(
	xfs_inode_t	*ip,
	xfs_off_t	first,
	xfs_off_t	last,
	uint64_t	flags,
	int		fiopt)
{
	struct address_space *mapping = VFS_I(ip)->i_mapping;
	int		ret = 0;

	if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
		xfs_iflags_clear(ip, XFS_ITRUNCATED);
		if (flags & XFS_B_ASYNC) {
			ret = __filemap_fdatawrite_range(mapping,
					0, LLONG_MAX, WB_SYNC_NONE);
		} else {
			int ret2;

			mapping_fsync_lock(mapping);
			ret = filemap_fdatawrite(mapping);
			ret2 = filemap_fdatawait_fsync(mapping);
			if (!ret)
				ret = ret2;
			mapping_fsync_unlock(mapping);
		}
	}
	return -ret;
}