From: Benny Halevy <benny@tonian.com>
Subject: Re: [PATCH] NFSv4.1: need to put_layout_hdr on _pnfs_return_layout
 error path
Date: Wed, 15 Jun 2011 11:44:02 -0400
Message-ID: <4DF8D342.6000700@tonian.com>
References: <1308152397-16920-1-git-send-email-benny@tonian.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Cc: Benny Halevy <benny@tonian.com>, linux-nfs@vger.kernel.org
To: trond.myklebust@netapp.com
In-Reply-To: <1308152397-16920-1-git-send-email-benny@tonian.com>
Sender: linux-nfs-owner@vger.kernel.org

On 2011-06-15 11:39, Benny Halevy wrote:
> We always get a reference on the layout header and we rely on
> nfs4_layoutreturn_release to put it.  If we hit an allocation error
> before starting the rpc proc we bail out early without dereferncing
> the layout header properly.
> 
> Signed-off-by: Benny Halevy <benny@tonian.com>
> ---
>  fs/nfs/nfs4proc.c |    1 +
>  fs/nfs/pnfs.c     |    1 +
>  2 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> index 79f3c33..a4705bc 100644
> --- a/fs/nfs/nfs4proc.c
> +++ b/fs/nfs/nfs4proc.c
> @@ -5774,6 +5774,7 @@ static void nfs4_layoutreturn_done(struct rpc_task *task, void *calldata)
>  	struct pnfs_layout_hdr *lo = NFS_I(lrp->args.inode)->layout;
>  
>  	dprintk("--> %s\n", __func__);
> +	dprintk("%s: ref %d\n", atonic_read(&lo->plh_refcount));
>  
>  	if (!nfs4_sequence_done(task, &lrp->res.seq_res))
>  		return;

please ignore this hunk,
got there totally by mistake...

Benny

> diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
> index 0feeccc..bc3eb74 100644
> --- a/fs/nfs/pnfs.c
> +++ b/fs/nfs/pnfs.c
> @@ -675,6 +675,7 @@ _pnfs_return_layout(struct inode *ino)
>  	lrp = kzalloc(sizeof(*lrp), GFP_KERNEL);
>  	if (unlikely(lrp == NULL)) {
>  		status = -ENOMEM;
> +		put_layout_hdr(lo);
>  		goto out;
>  	}
>