linux-kernel - Re: [PATCH v5 3/7] nfsd: only check RPC_SIGNALLED() when restarting rpc

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <610bc63d-1ce3-40d1-aa42-bfb22dc26ac4@oracle.com>
Date: Sat, 8 Feb 2025 11:59:28 -0500
From: Chuck Lever <chuck.lever@...cle.com>
To: Jeff Layton <jlayton@...nel.org>, Neil Brown <neilb@...e.de>,
        Olga Kornievskaia <okorniev@...hat.com>, Dai Ngo <Dai.Ngo@...cle.com>,
        Tom Talpey <tom@...pey.com>, "J. Bruce Fields" <bfields@...ldses.org>
Cc: linux-nfs@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5 3/7] nfsd: only check RPC_SIGNALLED() when restarting
 rpc_task

On 2/7/25 4:53 PM, Jeff Layton wrote:
> nfsd4_cb_sequence_done() currently checks RPC_SIGNALLED() when
> processing the compound and releasing the slot. If RPC_SIGNALLED()
> returns true, then that means that the client is going to be torn down.
> 
> Don't check RPC_SIGNALLED() after processing a successful reply.
> Instead, only check that before restarting the rpc_task. If that returns
> true, then requeue the callback.

This might seem like a nit, but this paragraph didn't make sense to me
at all until I changed s/only check that before/check that only before/.


> Also, handle rpc_restart_call() and rpc_restart_call_prepare() failures
> correctly, by requeueing the callback.
> 
> Signed-off-by: Jeff Layton <jlayton@...nel.org>
> ---
>  fs/nfsd/nfs4callback.c | 18 ++++++++++--------
>  1 file changed, 10 insertions(+), 8 deletions(-)
> 
> diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
> index bb5356e8713a8840bb714859618ff88130825efd..1e601075fac02bd5f01ff89d9252ab23d08c4fd9 100644
> --- a/fs/nfsd/nfs4callback.c
> +++ b/fs/nfsd/nfs4callback.c
> @@ -1385,8 +1385,8 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback
>  		goto requeue;
>  	case -NFS4ERR_DELAY:
>  		cb->cb_seq_status = 1;
> -		if (!rpc_restart_call(task))
> -			goto out;
> +		if (RPC_SIGNALLED(task) || !rpc_restart_call(task))
> +			goto requeue;
>  		rpc_delay(task, 2 * HZ);
>  		return false;
>  	case -NFS4ERR_BADSLOT:
> @@ -1402,14 +1402,16 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback
>  	}
>  	trace_nfsd_cb_free_slot(task, cb);
>  	nfsd41_cb_release_slot(cb);
> -
> -	if (RPC_SIGNALLED(task))
> -		goto requeue;
> -out:
>  	return ret;
>  retry_nowait:
> -	rpc_restart_call_prepare(task);
> -	goto out;
> +	/*
> +	 * RPC_SIGNALLED() means that the rpc_client is being torn down and
> +	 * (possibly) recreated. Requeue the call in that case.
> +	 */
> +	if (!RPC_SIGNALLED(task)) {
> +		if (rpc_restart_call_prepare(task))
> +			return false;
> +	}
>  requeue:
>  	nfsd41_cb_release_slot(cb);
>  	if (!test_bit(NFSD4_CLIENT_CB_KILL, &clp->cl_flags)) {
> 


-- 
Chuck Lever