xfs
[Top] [All Lists]

Re: [PATCH 05/16] xfs: decouple inode and bmap btree header files

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH 05/16] xfs: decouple inode and bmap btree header files
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Tue, 15 Oct 2013 12:06:20 -0700
Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20131004010009.GD4446@dastard>
References: <1380510433-8353-1-git-send-email-david@xxxxxxxxxxxxx> <1380510433-8353-6-git-send-email-david@xxxxxxxxxxxxx> <20131003145849.GE10316@xxxxxxxxxxxxx> <20131004010009.GD4446@dastard>
User-agent: Mutt/1.5.21 (2010-09-15)
On Fri, Oct 04, 2013 at 11:00:09AM +1000, Dave Chinner wrote:
> On Thu, Oct 03, 2013 at 07:58:49AM -0700, Christoph Hellwig wrote:
> > I like this a lot, but it also seems to move the in-core btree records
> > and keys to the format header which seems wrong.  Maybe those should be
> > centralized in xfs_btree.h instead?
> 
> Which structures are you talking about here? there's so many key and
> pointer definitions I'm not sure if there's a specific one you had
> in mind or whether you mean "all of them"...
> 
> Keep in mind here that the real dependency problem is the
> xfs_bmbt_rec_host definition being require by xfs_inode_fork.h.
> That's an in-core definition, and if we move that to another header
> file, we simply create a different dependency instead of removing
> the dependency altogether....

Exactly those.  It's been a while since that comment, so for now my
suggested approach would be to merge your series and try to clean things
up later.

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