[Top] [All Lists]

Re: [PATCH 1/7] xfs: don't dirty buffers beyond EOF

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH 1/7] xfs: don't dirty buffers beyond EOF
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Fri, 29 Aug 2014 10:53:58 +1000
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20140829003911.GA30463@xxxxxxxxxxxxx>
References: <1409226551-16570-1-git-send-email-david@xxxxxxxxxxxxx> <1409226551-16570-2-git-send-email-david@xxxxxxxxxxxxx> <20140829003911.GA30463@xxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Thu, Aug 28, 2014 at 05:39:11PM -0700, Christoph Hellwig wrote:
> Not sure if my 10 cents are worth anything given that I haven't spent
> much time with this code recently, but I feel very uneasy diverging
> from the generic path in this area.

I can't see how we have any other choice right now. We're caught
between a rock and a hard place - XFS uses bufferheads differently
to all other filesystems (esp. w.r.t. to EOF block zeroing
behaviour), and so changing behaviour in the generic code to suit
XFS is likely to introduce subtle data corruption bugs in other

I think the best thing we can do is move away from bufferheads in
XFS. We've already got lots of hacky code to manage
bufferhead/extent state coherency and so the sooner we get rid of
bufferheads the sooner that crap goes away, too.


Dave Chinner

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