Re: [PATCH net-next v4 1/3] net/smc: refactoring initialization of smc sock

[Tony Lu <tonylu@linux.alibaba.com>]

On Wed, May 29, 2024 at 11:59:05AM +0800, D. Wythe wrote:
> From: "D. Wythe" <alibuda@linux.alibaba.com>
> 
> This patch aims to isolate the shared components of SMC socket
> allocation by introducing smc_sk_init() for sock initialization
> and __smc_create_clcsk() for the initialization of clcsock.
> 
> This is in preparation for the subsequent implementation of the
> AF_INET version of SMC.
> 
> Signed-off-by: D. Wythe <alibuda@linux.alibaba.com>

v4 looks good for me, thanks.

Reviewed-by: Tony Lu <tonylu@linux.alibaba.com>

> ---
>  net/smc/af_smc.c | 86 +++++++++++++++++++++++++++++++-------------------------
>  net/smc/smc.h    |  5 ++++
>  2 files changed, 53 insertions(+), 38 deletions(-)
> 
> diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c
> index e50a286..77a9d58 100644
> --- a/net/smc/af_smc.c
> +++ b/net/smc/af_smc.c
> @@ -361,25 +361,15 @@ static void smc_destruct(struct sock *sk)
>  		return;
>  }
>  
> -static struct sock *smc_sock_alloc(struct net *net, struct socket *sock,
> -				   int protocol)
> +void smc_sk_init(struct net *net, struct sock *sk, int protocol)
>  {
> -	struct smc_sock *smc;
> -	struct proto *prot;
> -	struct sock *sk;
> -
> -	prot = (protocol == SMCPROTO_SMC6) ? &smc_proto6 : &smc_proto;
> -	sk = sk_alloc(net, PF_SMC, GFP_KERNEL, prot, 0);
> -	if (!sk)
> -		return NULL;
> +	struct smc_sock *smc = smc_sk(sk);
>  
> -	sock_init_data(sock, sk); /* sets sk_refcnt to 1 */
>  	sk->sk_state = SMC_INIT;
>  	sk->sk_destruct = smc_destruct;
>  	sk->sk_protocol = protocol;
>  	WRITE_ONCE(sk->sk_sndbuf, 2 * READ_ONCE(net->smc.sysctl_wmem));
>  	WRITE_ONCE(sk->sk_rcvbuf, 2 * READ_ONCE(net->smc.sysctl_rmem));
> -	smc = smc_sk(sk);
>  	INIT_WORK(&smc->tcp_listen_work, smc_tcp_listen_work);
>  	INIT_WORK(&smc->connect_work, smc_connect_work);
>  	INIT_DELAYED_WORK(&smc->conn.tx_work, smc_tx_work);
> @@ -389,6 +379,24 @@ static struct sock *smc_sock_alloc(struct net *net, struct socket *sock,
>  	sk->sk_prot->hash(sk);
>  	mutex_init(&smc->clcsock_release_lock);
>  	smc_init_saved_callbacks(smc);
> +	smc->limit_smc_hs = net->smc.limit_smc_hs;
> +	smc->use_fallback = false; /* assume rdma capability first */
> +	smc->fallback_rsn = 0;
> +}
> +
> +static struct sock *smc_sock_alloc(struct net *net, struct socket *sock,
> +				   int protocol)
> +{
> +	struct proto *prot;
> +	struct sock *sk;
> +
> +	prot = (protocol == SMCPROTO_SMC6) ? &smc_proto6 : &smc_proto;
> +	sk = sk_alloc(net, PF_SMC, GFP_KERNEL, prot, 0);
> +	if (!sk)
> +		return NULL;
> +
> +	sock_init_data(sock, sk); /* sets sk_refcnt to 1 */
> +	smc_sk_init(net, sk, protocol);
>  
>  	return sk;
>  }
> @@ -3321,6 +3329,31 @@ static ssize_t smc_splice_read(struct socket *sock, loff_t *ppos,
>  	.splice_read	= smc_splice_read,
>  };
>  
> +int smc_create_clcsk(struct net *net, struct sock *sk, int family)
> +{
> +	struct smc_sock *smc = smc_sk(sk);
> +	int rc;
> +
> +	rc = sock_create_kern(net, family, SOCK_STREAM, IPPROTO_TCP,
> +			      &smc->clcsock);
> +	if (rc) {
> +		sk_common_release(sk);
> +		return rc;
> +	}
> +
> +	/* smc_clcsock_release() does not wait smc->clcsock->sk's
> +	 * destruction;  its sk_state might not be TCP_CLOSE after
> +	 * smc->sk is close()d, and TCP timers can be fired later,
> +	 * which need net ref.
> +	 */
> +	sk = smc->clcsock->sk;
> +	__netns_tracker_free(net, &sk->ns_tracker, false);
> +	sk->sk_net_refcnt = 1;
> +	get_net_track(net, &sk->ns_tracker, GFP_KERNEL);
> +	sock_inuse_add(net, 1);
> +	return 0;
> +}
> +
>  static int __smc_create(struct net *net, struct socket *sock, int protocol,
>  			int kern, struct socket *clcsock)
>  {
> @@ -3346,35 +3379,12 @@ static int __smc_create(struct net *net, struct socket *sock, int protocol,
>  
>  	/* create internal TCP socket for CLC handshake and fallback */
>  	smc = smc_sk(sk);
> -	smc->use_fallback = false; /* assume rdma capability first */
> -	smc->fallback_rsn = 0;
> -
> -	/* default behavior from limit_smc_hs in every net namespace */
> -	smc->limit_smc_hs = net->smc.limit_smc_hs;
>  
>  	rc = 0;
> -	if (!clcsock) {
> -		rc = sock_create_kern(net, family, SOCK_STREAM, IPPROTO_TCP,
> -				      &smc->clcsock);
> -		if (rc) {
> -			sk_common_release(sk);
> -			goto out;
> -		}
> -
> -		/* smc_clcsock_release() does not wait smc->clcsock->sk's
> -		 * destruction;  its sk_state might not be TCP_CLOSE after
> -		 * smc->sk is close()d, and TCP timers can be fired later,
> -		 * which need net ref.
> -		 */
> -		sk = smc->clcsock->sk;
> -		__netns_tracker_free(net, &sk->ns_tracker, false);
> -		sk->sk_net_refcnt = 1;
> -		get_net_track(net, &sk->ns_tracker, GFP_KERNEL);
> -		sock_inuse_add(net, 1);
> -	} else {
> +	if (clcsock)
>  		smc->clcsock = clcsock;
> -	}
> -
> +	else
> +		rc = smc_create_clcsk(net, sk, family);
>  out:
>  	return rc;
>  }
> diff --git a/net/smc/smc.h b/net/smc/smc.h
> index 18c8b78..3edec1e 100644
> --- a/net/smc/smc.h
> +++ b/net/smc/smc.h
> @@ -34,6 +34,11 @@
>  extern struct proto smc_proto;
>  extern struct proto smc_proto6;
>  
> +/* smc sock initialization */
> +void smc_sk_init(struct net *net, struct sock *sk, int protocol);
> +/* clcsock initialization */
> +int smc_create_clcsk(struct net *net, struct sock *sk, int family);
> +
>  #ifdef ATOMIC64_INIT
>  #define KERNEL_HAS_ATOMIC64
>  #endif
> -- 
> 1.8.3.1