All of lore.kernel.org
 help / color / mirror / Atom feed
From: Trond Myklebust <Trond.Myklebust@netapp.com>
To: Alexandros Batsakis <batsakis@netapp.com>
Cc: linux-nfs@vger.kernel.org, trond@netapp.com
Subject: Re: [PATCH 3/4] nfs41: fix race between umount and renewd sequence operations
Date: Thu, 28 Jan 2010 15:47:46 -0500	[thread overview]
Message-ID: <1264711666.7553.45.camel@localhost> (raw)
In-Reply-To: <1264560235-2478-3-git-send-email-batsakis@netapp.com>

On Tue, 2010-01-26 at 18:43 -0800, Alexandros Batsakis wrote: 
> Signed-off-by: Alexandros Batsakis <batsakis@netapp.com>
> ---
>  fs/nfs/nfs4proc.c |   18 ++++++++++++++++--
>  1 files changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> index 9fc99e9..cd523df 100644
> --- a/fs/nfs/nfs4proc.c
> +++ b/fs/nfs/nfs4proc.c
> @@ -5041,6 +5041,8 @@ void nfs41_sequence_call_done(struct rpc_task *task, void *data)
>  
>  	if (task->tk_status < 0) {
>  		dprintk("%s ERROR %d\n", __func__, task->tk_status);
> +		if (atomic_read(&clp->cl_count) == 1)
> +			return;
>  
>  		if (_nfs4_async_handle_error(task, NULL, clp, NULL)
>  								== -EAGAIN) {
> @@ -5050,7 +5052,8 @@ void nfs41_sequence_call_done(struct rpc_task *task, void *data)
>  	}
>  	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
>  
> -	nfs4_schedule_state_renewal(clp);
> +	if (atomic_read(&clp->cl_count) > 1)

Why not just have a single
         if (atomic_read(&clp->cl_count) == 1)
return;

above

> +		nfs4_schedule_state_renewal(clp);
>  
>  	kfree(task->tk_msg.rpc_argp);
>  	kfree(task->tk_msg.rpc_resp);

These belong in the 'sequence_release' function regardless. Otherwise
you have a memory leak if rpc_call_async() failed, and also if your test
for clp->cl_count == 1 in the error case above.

> @@ -5073,9 +5076,15 @@ static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
>  	rpc_call_start(task);
>  }
>  
> +static void nfs41_sequence_release(void *calldata)
> +{
> +	nfs_put_client((struct nfs_client *) calldata);
> +}
> +
>  static const struct rpc_call_ops nfs41_sequence_ops = {
>  	.rpc_call_done = nfs41_sequence_call_done,
>  	.rpc_call_prepare = nfs41_sequence_prepare,
> +	.rpc_release = nfs41_sequence_release,
>  };
>  
>  static int nfs41_proc_async_sequence(struct nfs_client *clp,
> @@ -5088,11 +5097,16 @@ static int nfs41_proc_async_sequence(struct nfs_client *clp,
>  		.rpc_cred = cred,
>  	};
>  
> +	if (!atomic_inc_not_zero(&clp->cl_count))
> +		return -EIO;
>  	args = kzalloc(sizeof(*args), GFP_KERNEL);
> -	if (!args)
> +	if (!args) {
> +		nfs_put_client(clp);
>  		return -ENOMEM;
> +	}
>  	res = kzalloc(sizeof(*res), GFP_KERNEL);
>  	if (!res) {
> +		nfs_put_client(clp);
>  		kfree(args);
>  		return -ENOMEM;
>  	}



      parent reply	other threads:[~2010-01-28 20:48 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-27  2:43 [PATCH 1/4] nfs: kill renewd before clearing the client session Alexandros Batsakis
2010-01-27  2:43 ` [PATCH 2/4] nfs4: prevent backlogging of renewd requests Alexandros Batsakis
2010-01-27  2:43   ` [PATCH 3/4] nfs41: fix race between umount and renewd sequence operations Alexandros Batsakis
2010-01-27  2:43     ` [PATCH 4/4] nfs4: fix race between umount and renew operations Alexandros Batsakis
2010-01-28 20:48       ` Trond Myklebust
2010-01-28 20:47     ` Trond Myklebust [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1264711666.7553.45.camel@localhost \
    --to=trond.myklebust@netapp.com \
    --cc=batsakis@netapp.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=trond@netapp.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.