Re: [PATCH] smc: fix refcount bug in sk_psock_get (2)

From: Jakub Kicinski <kuba@kernel.org>
To: Hawkins Jiawei <yin31149@gmail.com>
Cc: syzbot+5f26f85569bd179c18ce@syzkaller.appspotmail.com,
	andrii@kernel.org, ast@kernel.org, borisp@nvidia.com,
	bpf@vger.kernel.org, daniel@iogearbox.net, davem@davemloft.net,
	edumazet@google.com, guwen@linux.alibaba.com,
	yoshfuji@linux-ipv6.org, dsahern@kernel.org,
	john.fastabend@gmail.com, kafai@fb.com, kgraul@linux.ibm.com,
	kpsingh@kernel.org, linux-kernel@vger.kernel.org,
	netdev@vger.kernel.org, pabeni@redhat.com, songliubraving@fb.com,
	syzkaller-bugs@googlegroups.com, yhs@fb.com,
	skhan@linuxfoundation.org, 18801353760@163.com,
	paskripkin@gmail.com,
	linux-kernel-mentees@lists.linuxfoundation.org
Subject: Re: [PATCH] smc: fix refcount bug in sk_psock_get (2)
Date: Fri, 8 Jul 2022 20:06:02 -0700	[thread overview]
Message-ID: <20220708200602.1059bc09@kernel.org> (raw)
In-Reply-To: <20220709024659.6671-1-yin31149@gmail.com>

On Sat,  9 Jul 2022 10:46:59 +0800 Hawkins Jiawei wrote:
> Reported-and-tested-by: syzbot+5f26f85569bd179c18ce@syzkaller.appspotmail.com
> Signed-off-by: hawk <18801353760@163.com>
> ---
>  net/ipv4/tcp.c | 13 +++++++++++++
>  1 file changed, 13 insertions(+)
> 
> diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> index 9984d23a7f3e..a1e6cab2c748 100644
> --- a/net/ipv4/tcp.c
> +++ b/net/ipv4/tcp.c
> @@ -3395,10 +3395,23 @@ static int do_tcp_setsockopt(struct sock *sk, int level, int optname,
>  	}
>  	case TCP_ULP: {
>  		char name[TCP_ULP_NAME_MAX];
> +		struct sock *smc_sock;
>  
>  		if (optlen < 1)
>  			return -EINVAL;
>  
> +		/* SMC sk_user_data may be treated as psock,
> +		 * which triggers a refcnt warning.
> +		 */
> +		rcu_read_lock();
> +		smc_sock = rcu_dereference_sk_user_data(sk);
> +		if (level == SOL_TCP && smc_sock &&
> +		    smc_sock->__sk_common.skc_family == AF_SMC) {

This should prolly be under the socket lock?

Can we add a bit to SK_USER_DATA_PTRMASK and have ULP-compatible
users (sockmap) opt into ULP cooperation? Modifying TCP is backwards,
layer-wise.

> +			rcu_read_unlock();
> +			return -EOPNOTSUPP;
> +		}
> +		rcu_read_unlock();
> +