From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cuda.sgi.com (cuda1.sgi.com [192.48.157.11]) by oss.sgi.com (8.14.3/8.14.3/SuSE Linux 0.8) with ESMTP id q92KNY1m083696 for ; Tue, 2 Oct 2012 15:23:34 -0500 Received: from ipmail04.adl6.internode.on.net (ipmail04.adl6.internode.on.net [150.101.137.141]) by cuda.sgi.com with ESMTP id mKNIJFAtVa4bLr8F for ; Tue, 02 Oct 2012 13:24:57 -0700 (PDT) Date: Wed, 3 Oct 2012 06:24:56 +1000 From: Dave Chinner Subject: Re: [PATCH 06/13] xfs: xfs_sync_data is redundant. Message-ID: <20121002202456.GR23520@dastard> References: <1348807485-20165-1-git-send-email-david@fromorbit.com> <1348807485-20165-7-git-send-email-david@fromorbit.com> <5069F9B0.50804@redhat.com> <20121002001021.GJ23520@dastard> <20121002132205.GA14572@infradead.org> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <20121002132205.GA14572@infradead.org> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: xfs-bounces@oss.sgi.com Errors-To: xfs-bounces@oss.sgi.com To: Christoph Hellwig Cc: Brian Foster , xfs@oss.sgi.com On Tue, Oct 02, 2012 at 09:22:05AM -0400, Christoph Hellwig wrote: > On Tue, Oct 02, 2012 at 10:10:22AM +1000, Dave Chinner wrote: > > diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h > > index da69c18..0ec7a46 100644 > > --- a/fs/xfs/xfs_inode.h > > +++ b/fs/xfs/xfs_inode.h > > @@ -294,7 +294,7 @@ xfs_new_eof(struct xfs_inode *ip, xfs_fsize_t new_size) > > static inline void > > xfs_flush_inodes(struct xfs_inode *ip) > > { > > - writeback_inodes_sb_if_idle(VFS_I(ip)->i_sb, WB_REASON_FS_FREE_SPACE); > > + sync_inodes_sb(VFS_I(ip)->i_sb); > > sync_inodes_sb needs s_umount held and asserts that, while our callers > usually won't have it. Ah, bugger. I didn't notice that, and hadn't checked dmesg after I ran 273 a few times with this patch. I just figured that because it waited for completion, it didn't need it. Adding s_umount is fine for the wirte path, but the create path holds a directory i_mutex so I'd guess that means it has to be a trylock to avoid lock inversion warnings from lockdep.... Cheers, Dave. -- Dave Chinner david@fromorbit.com _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs