xfs
[Top] [All Lists]

Re: [PATCH 1/2] fsx: Add fallocate collapse range operation

To: LukÃÅ Czerner <lczerner@xxxxxxxxxx>
Subject: Re: [PATCH 1/2] fsx: Add fallocate collapse range operation
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Fri, 4 Apr 2014 08:26:37 +1100
Cc: xfs@xxxxxxxxxxx, linux-ext4@xxxxxxxxxxxxxxx, linux-fsdevel@xxxxxxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <alpine.LFD.2.00.1404021752340.2169@xxxxxxxxxxxxxxxxxxxxx>
References: <1396452800-25775-1-git-send-email-lczerner@xxxxxxxxxx> <alpine.LFD.2.00.1404021752340.2169@xxxxxxxxxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Wed, Apr 02, 2014 at 05:55:07PM +0200, LukÃÅ Czerner wrote:
> On Wed, 2 Apr 2014, Lukas Czerner wrote:
> 
> > Date: Wed,  2 Apr 2014 17:33:19 +0200
> > From: Lukas Czerner <lczerner@xxxxxxxxxx>
> > To: xfs@xxxxxxxxxxx
> > Cc: linux-ext4@xxxxxxxxxxxxxxx, linux-fsdevel@xxxxxxxxxxxxxxx,
> >     Lukas Czerner <lczerner@xxxxxxxxxx>
> > Subject: [PATCH 1/2] fsx: Add fallocate collapse range operation
> > 
> > This commit adds fallocate FALLOC_FL_COLLAPSE_RANGE support for fsx.
> 
> Btw, this actually makes the fsx fail on ext4 and xfs after a while.
> I was trying to find a problem in fsx itself but I think it is
> actually ok. That means that we could possibly have bugs in ext4 and
> xfs collapse range implementation. More eyes are needed on this!

No surprise, really. As it is, your previous ZERO_RANGE additions
caused problems for XFS, and those are mostly fixed in the 6 patch
series I sent before LSFMM.  Hence I wouldn't be surprised if that's
what you are seeing (on XFs, at least) and it's not actually
COLLAPSE_RANGE that is causing problems...

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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