From mboxrd@z Thu Jan  1 00:00:00 1970
From: Chris Mason <chris.mason@oracle.com>
Subject: Re: Algorithm for nodatacow is broken
Date: Wed, 16 Jul 2008 13:15:54 +0000
Message-ID: <1216214154.24425.326.camel@think.oraclecorp.com>
References: <3d0408630807160041h20d2157agfc221b34a836df87@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain
Cc: linux-btrfs@vger.kernel.org
To: Yan Zheng <yanzheng@21cn.com>
Return-path: <linux-btrfs-owner@vger.kernel.org>
In-Reply-To: <3d0408630807160041h20d2157agfc221b34a836df87@mail.gmail.com>
List-ID: <linux-btrfs.vger.kernel.org>

On Wed, 2008-07-16 at 15:41 +0800, Yan Zheng wrote:
> Hello,
> 
> Yesterday, I realized the algorithm for nodatacow is broken, it can't
> reliably detect whether a given extent is referenced by only one
> snapshot.

I had to change around nodatacow back in May because it was definitely
broken in the way you describe.  I agree the special case I added to
allow nodatacow when one of the references comes from the running
transaction is broken.

But, we should be able to fix it by extending the reference count checks
up the tree.  Any reference > 1 not held by the running transaction on
any block should force a cow.

-chris