Received-SPF: pass (google.com: best guess record for domain of linux-nfs-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Message-ID: <cb9b5115fc445a5d775b6cdb6f58917e7c80b154.camel@gmail.com>
Subject: Re: [PATCH 4/6] NFS: Have nfs41_proc_reclaim_complete() call
 nfs4_call_sync_custom()
From:   Anna Schumaker <schumaker.anna@gmail.com>
To:     Trond Myklebust <trondmy@hammerspace.com>,
        "linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>
Cc:     NeilBrown <neilb@suse.com>
Date:   Mon, 19 Aug 2019 16:02:32 -0400
In-Reply-To: <8bd34fcbd352a2d5c4a8c757919f044bfaa76c60.camel@hammerspace.com>
References: <20190819192900.19312-1-Anna.Schumaker@Netapp.com>
         <20190819192900.19312-5-Anna.Schumaker@Netapp.com>
         <8bd34fcbd352a2d5c4a8c757919f044bfaa76c60.camel@hammerspace.com>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.32.4 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-nfs-owner@vger.kernel.org
Precedence: bulk

See 5a0c257f8e0f4c4b3c33dff545317c21a921303e (NFS: send state
management on a single connection)

My understanding is that it forces all the state management calls into
a single connection during session startup, but before the extra
network connections are bound to the session.

Anna

On Mon, 2019-08-19 at 19:56 +0000, Trond Myklebust wrote:
> On Mon, 2019-08-19 at 15:28 -0400, schumaker.anna@gmail.com wrote:
> > From: Anna Schumaker <Anna.Schumaker@Netapp.com>
> > 
> > An async call followed by an rpc_wait_for_completion() is basically
> > the
> > same as a synchronous call, so we can use nfs4_call_sync_custom()
> > to
> > keep our custom callback ops and the RPC_TASK_NO_ROUND_ROBIN flag.
> > 
> > Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
> > ---
> >  fs/nfs/nfs4proc.c | 13 ++-----------
> >  1 file changed, 2 insertions(+), 11 deletions(-)
> > 
> > diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
> > index de2b3fd806ef..1b7863ec12d3 100644
> > --- a/fs/nfs/nfs4proc.c
> > +++ b/fs/nfs/nfs4proc.c
> > @@ -8857,7 +8857,6 @@ static int nfs41_proc_reclaim_complete(struct
> > nfs_client *clp,
> >  		const struct cred *cred)
> >  {
> >  	struct nfs4_reclaim_complete_data *calldata;
> > -	struct rpc_task *task;
> >  	struct rpc_message msg = {
> >  		.rpc_proc =
> > &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
> >  		.rpc_cred = cred,
> > @@ -8866,7 +8865,7 @@ static int nfs41_proc_reclaim_complete(struct
> > nfs_client *clp,
> >  		.rpc_client = clp->cl_rpcclient,
> >  		.rpc_message = &msg,
> >  		.callback_ops = &nfs4_reclaim_complete_call_ops,
> > -		.flags = RPC_TASK_ASYNC | RPC_TASK_NO_ROUND_ROBIN,
> > +		.flags = RPC_TASK_NO_ROUND_ROBIN,
> >  	};
> >  	int status = -ENOMEM;
> >  
> > @@ -8881,15 +8880,7 @@ static int
> > nfs41_proc_reclaim_complete(struct
> > nfs_client *clp,
> >  	msg.rpc_argp = &calldata->arg;
> >  	msg.rpc_resp = &calldata->res;
> >  	task_setup_data.callback_data = calldata;
> > -	task = rpc_run_task(&task_setup_data);
> > -	if (IS_ERR(task)) {
> > -		status = PTR_ERR(task);
> > -		goto out;
> > -	}
> > -	status = rpc_wait_for_completion_task(task);
> > -	if (status == 0)
> > -		status = task->tk_status;
> > -	rpc_put_task(task);
> > +	status = nfs4_call_sync_custom(&task_setup_data);
> >  out:
> >  	dprintk("<-- %s status=%d\n", __func__, status);
> >  	return status;
> 
> Hmm... I'm a little confused. Why does RECLAIM_COMPLETE need
> RPC_TASK_NO_ROUND_ROBIN? It should be ordered so it is called after
> BIND_CONN_TO_SESSION in nfs4_state_manager(), so in principle it is
> supposed to be able to recover from an error like
> NFS4ERR_CONN_NOT_BOUND_TO_SESSION. Are there other situations where
> we
> need RPC_TASK_NO_ROUND_ROBIN?
> 
> -- 
> Trond Myklebust
> Linux NFS client maintainer, Hammerspace
> trond.myklebust@hammerspace.com
> 
>