xfs
[Top] [All Lists]

Re: [PATCH 0/12 v2] Moving i_dquot out of struct inode

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH 0/12 v2] Moving i_dquot out of struct inode
From: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Date: Sun, 12 Oct 2014 19:53:16 +0100
Cc: Jan Kara <jack@xxxxxxx>, linux-fsdevel@xxxxxxxxxxxxxxx, linux-ext4@xxxxxxxxxxxxxxx, Dave Chinner <david@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx, cluster-devel@xxxxxxxxxx, Steven Whitehouse <swhiteho@xxxxxxxxxx>, Mark Fasheh <mfasheh@xxxxxxxx>, Joel Becker <jlbec@xxxxxxxxxxxx>, ocfs2-devel@xxxxxxxxxxxxxx, reiserfs-devel@xxxxxxxxxxxxxxx, Jeff Mahoney <jeffm@xxxxxxx>, Dave Kleikamp <shaggy@xxxxxxxxxx>, jfs-discussion@xxxxxxxxxxxxxxxxxxxxx, tytso@xxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20141011133452.GA29004@xxxxxxxxxxxxx>
References: <1412952910-7142-1-git-send-email-jack@xxxxxxx> <20141011133452.GA29004@xxxxxxxxxxxxx>
Sender: Al Viro <viro@xxxxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Sat, Oct 11, 2014 at 06:34:52AM -0700, Christoph Hellwig wrote:
> I still very much disagree with the s_inode_fields indirection.  Please
> find a patch below to remove it, and use a get_dquots super_block
> operation instead.  This leads to less and better readable code,
> and serves 4 bytes in every inode in the system.  Additionally the
> indirection could easily be optimized away by directly passing the
> dquot array in various functions, but for now I'd like to keep it
> simple.

Indeed.  This "array of offsets" approach is asking for trouble.  Please,
don't go there - playing that way with type safety is a bad idea.

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