From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from out30-112.freemail.mail.aliyun.com (out30-112.freemail.mail.aliyun.com [115.124.30.112]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 2DD0A2BD02A; Tue, 10 Mar 2026 05:51:15 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=115.124.30.112 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1773121886; cv=none; b=IWhUf1QxRw5us6pPYH6vWM/oHwk+YuruMQ6HVr2Tv2zKwp2aEKfB2J+P8h5161siLQ/bq8iuqYjqwT1GE/DQj41JLXu3VbUNNjMokBSKkymxcb90T4BxrO46HZw8greLCItEID1EK2NxeDrsU2E0dGROgyPizbErmWlyq1qJDdI= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1773121886; c=relaxed/simple; bh=X8GlQBtv2iRJp4SsZ7G2UVyw+BbphQ1aA83k4nF9d6c=; h=Date:From:To:Cc:Subject:Message-ID:References:MIME-Version: Content-Type:Content-Disposition:In-Reply-To; b=XKqyne5bNixO4MzhO12lGTxvKzLdlMnkD5lzrwzPOcQFewJs0GUwu9xA2V2F2uW8OHB5y84X9HaFcSQ5vkWqJ3rOx7Ik3/+65YVTcH1sYVZkBvfivsb1yeR5XSAL/SpOOmmPOzbegvFCyNjNyLEGhjGj48iTNiMpefT6i4Rpzg0= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com; spf=pass smtp.mailfrom=linux.alibaba.com; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b=WB6SmFvI; arc=none smtp.client-ip=115.124.30.112 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b="WB6SmFvI" DKIM-Signature:v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.alibaba.com; s=default; t=1773121872; h=Date:From:To:Subject:Message-ID:MIME-Version:Content-Type; bh=RmK1Nc9lCr9lQbUEEr8YJz/ZJtYUzCfVYaSHJv4uhls=; b=WB6SmFvIJeafk29YQHGaysSZvN6NpuoLIvqMIH7p+ZXqDbpO88yzpP+rB1HKYHNka5aF/iHsx1LI/TL1+c6aZ9Z2DQaSwSLgKGXV6Ka75eHfhe31JDe36fvyL4OSX1qcHxN6K1YF3saUUhWTp75hYZC7FBZp0bIs8ZOfvxXU//k= Received: from localhost(mailfrom:alibuda@linux.alibaba.com fp:SMTPD_---0X-f-NWz_1773121871 cluster:ay36) by smtp.aliyun-inc.com; Tue, 10 Mar 2026 13:51:11 +0800 Date: Tue, 10 Mar 2026 13:51:11 +0800 From: "D. Wythe" To: Jiayuan Chen Cc: "D. Wythe" , netdev@vger.kernel.org, syzbot+827ae2bfb3a3529333e9@syzkaller.appspotmail.com, Eric Dumazet , Dust Li , Sidraya Jayagond , Wenjia Zhang , Mahanta Jambigi , Tony Lu , Wen Gu , "David S. Miller" , Jakub Kicinski , Paolo Abeni , Simon Horman , linux-rdma@vger.kernel.org, linux-s390@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH net v2] net/smc: fix NULL dereference and UAF in smc_tcp_syn_recv_sock() Message-ID: <20260310055111.GB40558@j66a10360.sqa.eu95> References: <20260309023846.18516-1-jiayuan.chen@linux.dev> <20260309060611.GA130186@j66a10360.sqa.eu95> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) On Mon, Mar 09, 2026 at 07:48:54AM +0000, Jiayuan Chen wrote: > March 9, 2026 at 14:06, "D. Wythe" wrote: > > > [...] > > > Reproducer was verified with mdelay injection and smc_run, > > > the issue no longer occurs with this patch applied. > > > > > > [1] https://syzkaller.appspot.com/bug?extid=827ae2bfb3a3529333e9 > > > > > > Fixes: 8270d9c21041 ("net/smc: Limit backlog connections") > > > Reported-by: syzbot+827ae2bfb3a3529333e9@syzkaller.appspotmail.com > > > Closes: https://lore.kernel.org/all/67eaf9b8.050a0220.3c3d88.004a.GAE@google.com/T/ > > > Suggested-by: Eric Dumazet > > > Signed-off-by: Jiayuan Chen > > > --- > > > v2: > > > - Use rcu_read_lock() + refcount_inc_not_zero() instead of > > > read_lock_bh(sk_callback_lock) + sock_hold(), since this > > > is the TCP handshake hot path and read_lock_bh is too > > > expensive under SYN flood. > > > - Set SOCK_RCU_FREE on SMC listen socket to ensure > > > RCU-deferred freeing. > > > > > > v1: https://lore.kernel.org/netdev/20260307032158.372165-1-jiayuan.chen@linux.dev/ > > > --- > > > net/smc/af_smc.c | 9 +++++++++ > > > 1 file changed, 9 insertions(+) > > > > > > diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c > > > index d0119afcc6a1..72ac1d8c62d4 100644 > > > --- a/net/smc/af_smc.c > > > +++ b/net/smc/af_smc.c > > > @@ -131,7 +131,13 @@ static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk, > > > struct smc_sock *smc; > > > struct sock *child; > > > > > > + rcu_read_lock(); > > > smc = smc_clcsock_user_data(sk); > > > + if (!smc || !refcount_inc_not_zero(&smc->sk.sk_refcnt)) { > > > + rcu_read_unlock(); > > > + return NULL; > > > + } > > > + rcu_read_unlock(); > > > > > > if (READ_ONCE(sk->sk_ack_backlog) + atomic_read(&smc->queued_smc_hs) > > > > sk->sk_max_ack_backlog) > > > @@ -153,11 +159,13 @@ static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk, > > > if (inet_csk(child)->icsk_af_ops == inet_csk(sk)->icsk_af_ops) > > > inet_csk(child)->icsk_af_ops = smc->ori_af_ops; > > > } > > > + sock_put(&smc->sk); > > > return child; > > > > > > drop: > > > dst_release(dst); > > > tcp_listendrop(sk); > > > + sock_put(&smc->sk); > > > return NULL; > > > } > > > > > > @@ -2691,6 +2699,7 @@ int smc_listen(struct socket *sock, int backlog) > > > write_unlock_bh(&smc->clcsock->sk->sk_callback_lock); > > > goto out; > > > } > > > + sock_set_flag(sk, SOCK_RCU_FREE); > > > > > This RCU approach looks good to me. Since SOCK_RCU_FREE is now enabled, > > other callers of smc_clcsock_user_data() should also follow this > > RCU-based pattern. It will eventually allow us to completely remove the > > annoying sk_callback_lock. > > > > D. Wythe > > > > > Hi D. Wythe, > > Thanks for the suggestion. I agree that converting all smc_clcsock_user_data() > callers to RCU is a reasonable direction, and it would allow us to eventually > remove the sk_callback_lock dependency for sk_user_data access. > > However, I'd prefer to keep this patch focused on fixing the specific bug in > smc_tcp_syn_recv_sock(), since it needs to be backported to stable trees. > Mixing a bug fix with broader refactoring makes backporting harder and increases > the risk of regressions. > > Also, converting the other callers is not entirely trivial. For example: > > - smc_fback_state_change/data_ready/write_space/error_report(): > the sk_callback_lock there protects not only sk_user_data but also the > consistency of the saved callback pointers (e.g.,smc->clcsk_state_change). > Switching to RCU requires careful ordering analysis against the write side > in smc_fback_restore_callbacks(). Additionally, fallback sockets would also > need SOCK_RCU_FREE. > > - smc_clcsock_data_ready(): > the sock_hold() would need to become refcount_inc_not_zero() to handle the > case where the refcount has already reached zero. > > I'd like to address these conversions in a follow-up patch series. > > What do you think? Thanks for the detailed explanation. I agree with your plan, it's better to keep the bug fix and the optimization separate. > > > > > > > sk->sk_max_ack_backlog = backlog; > > > sk->sk_ack_backlog = 0; > > > sk->sk_state = SMC_LISTEN; > > > -- > > > 2.43.0 > > > > >