xfs
[Top] [All Lists]

Re: [PATCH] XFS: Don't flush stale inodes

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH] XFS: Don't flush stale inodes
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Sun, 3 Jan 2010 00:39:02 +1100
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20100102131701.GA17320@xxxxxxxxxxxxx>
References: <1262399980-19277-1-git-send-email-david@xxxxxxxxxxxxx> <20100102120053.GB18502@xxxxxxxxxxxxx> <20100102122405.GI13802@xxxxxxxxxxxxxxxx> <20100102131701.GA17320@xxxxxxxxxxxxx>
User-agent: Mutt/1.5.18 (2008-05-17)
On Sat, Jan 02, 2010 at 08:17:01AM -0500, Christoph Hellwig wrote:
> This looks like a bigger problem to me.  We only mark inodes as stale
> from xfs_ifree_cluster, which via xfs_ifree and xfs_inactive gets
> called from xfs_fs_clear_inode.  Given that the inode has now been
> deleted we should not mark it as reclaimable in xfs_fs_destroy_inode
> but go on to reap it given that there is nothing to reclaim.

We can't reap it immediately as the inodes are usually dirty and
pinned in memory at this point by transactions that have not yet
been written to disk and completed. We need the inodes in core for
log IO completion processing....

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

<Prev in Thread] Current Thread [Next in Thread>