xfs
[Top] [All Lists]

Re: [PATCH 0/2] Add FALLOC_FL_ZERO_RANGE to fallocate

To: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Subject: Re: [PATCH 0/2] Add FALLOC_FL_ZERO_RANGE to fallocate
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Wed, 13 Jun 2012 13:30:52 +1000
Cc: linux-kernel@xxxxxxxxxxxxxxx, linux-fsdevel@xxxxxxxxxxxxxxx, Al Viro <viro@xxxxxxxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
In-reply-to: <20120613013549.GO22848@dastard>
References: <1339515364-17374-1-git-send-email-pbonzini@xxxxxxxxxx> <20120613013549.GO22848@dastard>
User-agent: Mutt/1.5.21 (2010-09-15)
On Wed, Jun 13, 2012 at 11:35:49AM +1000, Dave Chinner wrote:
> On Tue, Jun 12, 2012 at 05:36:02PM +0200, Paolo Bonzini wrote:
> > This patch adds the FALLOC_FL_ZERO_RANGE operation mode to fallocate,
> > resembling the similar XFS ioctl.  The new mode can be used with
> > or without FALLOC_FL_KEEP_SIZE, but of course not together with
> > FALLOC_FL_PUNCH_HOLE.
> 
> !FALLOC_FL_KEEP_SIZE makes no sense for this operation. It is for
> zeroing an existing section of a file while retaining the allocated
> space, not for extending or truncating the file. It's the same
> reason that FALLOC_FL_PUNCH_HOLE must have FALLOC_FL_KEEP_SIZE set.
> 
> Also, a minor nit, but you should credit where this code has
> originated from in the commit messages, and describe the use case
> for requiring it. i.e. based on:
> 
> http://permalink.gmane.org/gmane.linux.file-systems/62449

Oh, and you'll need to provide a new xfstests test based on 242 that
uses the new fallocate interface, and to do that you'll also need to
add support for the new fallocate function in xfs_io.

Adding generic filesystem functionality is not as simple as writing a
kernel patch anymore... :/

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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