From patchwork Mon Dec 11 10:29:23 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Davide Caratti X-Patchwork-Id: 13486990 X-Patchwork-Delegate: matthieu.baerts@tessares.net Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (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 9A43E24B33 for ; Mon, 11 Dec 2023 10:29:38 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="ilhF/6cv" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1702290577; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding; bh=1napOYuMOQmorrMXNoIH+dLhPjVi4RCCUPbnDZWHKOY=; b=ilhF/6cv1cHv+yJWZEMex2Y7yn+V0igmxQA69KAGRwDHhW6m1QfTivkCEu4XbyMjodK1Kf gHVRvP7yhbAy+y/NVyRuJvwIHdFHITmmmpH/gd1s+VUmV1JaDXxgJZjAkP6GST6FmZ8OIr B/rSaYS1y40YXRBHZ0L3GubiY6cic4Q= Received: from mimecast-mx02.redhat.com (mimecast-mx02.redhat.com [66.187.233.88]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-677-rRFQBMfeMvuTygzTiuMgGg-1; Mon, 11 Dec 2023 05:29:35 -0500 X-MC-Unique: rRFQBMfeMvuTygzTiuMgGg-1 Received: from smtp.corp.redhat.com (int-mx01.intmail.prod.int.rdu2.redhat.com [10.11.54.1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id B5EF4101A551 for ; Mon, 11 Dec 2023 10:29:35 +0000 (UTC) Received: from dcaratti.users.ipa.redhat.com (unknown [10.45.225.183]) by smtp.corp.redhat.com (Postfix) with ESMTP id 504FB3C25 for ; Mon, 11 Dec 2023 10:29:35 +0000 (UTC) From: Davide Caratti To: mptcp@lists.linux.dev Subject: [PATCH mptcp-next v2] mptcp: don't overwrite sock_ops in mptcp_is_tcpsk() Date: Mon, 11 Dec 2023 11:29:23 +0100 Message-ID: <784668db3e21351c91f4f312720059c6fa5cc15d.1702290281.git.dcaratti@redhat.com> Precedence: bulk X-Mailing-List: mptcp@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.4.1 on 10.11.54.1 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Eric Dumazet suggests: > The fact that mptcp_is_tcpsk() was able to write over sock->ops was a > bit strange to me. > mptcp_is_tcpsk() should answer a question, with a read-only argument. re-factor code to avoid overwriting sock_ops inside that function. Also, change the helper name to reflect the semantics and to disambiguate from its dual, sk_is_mptcp(). While at it, collapse mptcp_stream_accept() and mptcp_accept() into a single function, where fallback / non-fallback are separated into a single sk_is_mptcp() conditional. Link: https://github.com/multipath-tcp/mptcp_net-next/issues/432 Suggested-by: Eric Dumazet Signed-off-by: Davide Caratti Acked-by: Paolo Abeni --- net/mptcp/protocol.c | 108 ++++++++++++++++++------------------------- 1 file changed, 44 insertions(+), 64 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 4fc038d29623..b7cb60e6f0c3 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -55,28 +55,14 @@ u64 mptcp_wnd_end(const struct mptcp_sock *msk) return READ_ONCE(msk->wnd_end); } -static bool mptcp_is_tcpsk(struct sock *sk) +static const struct proto_ops *mptcp_fallback_tcp_ops(const struct sock *sk) { - struct socket *sock = sk->sk_socket; - - if (unlikely(sk->sk_prot == &tcp_prot)) { - /* we are being invoked after mptcp_accept() has - * accepted a non-mp-capable flow: sk is a tcp_sk, - * not an mptcp one. - * - * Hand the socket over to tcp so all further socket ops - * bypass mptcp. - */ - WRITE_ONCE(sock->ops, &inet_stream_ops); - return true; #if IS_ENABLED(CONFIG_MPTCP_IPV6) - } else if (unlikely(sk->sk_prot == &tcpv6_prot)) { - WRITE_ONCE(sock->ops, &inet6_stream_ops); - return true; + if (sk->sk_prot == &tcpv6_prot) + return &inet6_stream_ops; #endif - } - - return false; + WARN_ON_ONCE(sk->sk_prot != &tcp_prot); + return &inet_stream_ops; } static int __mptcp_socket_create(struct mptcp_sock *msk) @@ -3258,44 +3244,6 @@ void mptcp_rcv_space_init(struct mptcp_sock *msk, const struct sock *ssk) WRITE_ONCE(msk->wnd_end, msk->snd_nxt + tcp_sk(ssk)->snd_wnd); } -static struct sock *mptcp_accept(struct sock *ssk, int flags, int *err, - bool kern) -{ - struct sock *newsk; - - pr_debug("ssk=%p, listener=%p", ssk, mptcp_subflow_ctx(ssk)); - newsk = inet_csk_accept(ssk, flags, err, kern); - if (!newsk) - return NULL; - - pr_debug("newsk=%p, subflow is mptcp=%d", newsk, sk_is_mptcp(newsk)); - if (sk_is_mptcp(newsk)) { - struct mptcp_subflow_context *subflow; - struct sock *new_mptcp_sock; - - subflow = mptcp_subflow_ctx(newsk); - new_mptcp_sock = subflow->conn; - - /* is_mptcp should be false if subflow->conn is missing, see - * subflow_syn_recv_sock() - */ - if (WARN_ON_ONCE(!new_mptcp_sock)) { - tcp_sk(newsk)->is_mptcp = 0; - goto out; - } - - newsk = new_mptcp_sock; - MPTCP_INC_STATS(sock_net(ssk), MPTCP_MIB_MPCAPABLEPASSIVEACK); - } else { - MPTCP_INC_STATS(sock_net(ssk), - MPTCP_MIB_MPCAPABLEPASSIVEFALLBACK); - } - -out: - newsk->sk_kern_sock = kern; - return newsk; -} - void mptcp_destroy_common(struct mptcp_sock *msk, unsigned int flags) { struct mptcp_subflow_context *subflow, *tmp; @@ -3738,7 +3686,6 @@ static struct proto mptcp_prot = { .connect = mptcp_connect, .disconnect = mptcp_disconnect, .close = mptcp_close, - .accept = mptcp_accept, .setsockopt = mptcp_setsockopt, .getsockopt = mptcp_getsockopt, .shutdown = mptcp_shutdown, @@ -3848,18 +3795,36 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock, if (!ssk) return -EINVAL; - newsk = mptcp_accept(ssk, flags, &err, kern); + pr_debug("ssk=%p, listener=%p", ssk, mptcp_subflow_ctx(ssk)); + newsk = inet_csk_accept(ssk, flags, &err, kern); if (!newsk) return err; - lock_sock(newsk); - - __inet_accept(sock, newsock, newsk); - if (!mptcp_is_tcpsk(newsock->sk)) { - struct mptcp_sock *msk = mptcp_sk(newsk); + pr_debug("newsk=%p, subflow is mptcp=%d", newsk, sk_is_mptcp(newsk)); + if (sk_is_mptcp(newsk)) { struct mptcp_subflow_context *subflow; + struct sock *new_mptcp_sock; + + subflow = mptcp_subflow_ctx(newsk); + new_mptcp_sock = subflow->conn; + + /* is_mptcp should be false if subflow->conn is missing, see + * subflow_syn_recv_sock() + */ + if (WARN_ON_ONCE(!new_mptcp_sock)) { + tcp_sk(newsk)->is_mptcp = 0; + goto tcpfallback; + } + + newsk = new_mptcp_sock; + MPTCP_INC_STATS(sock_net(ssk), MPTCP_MIB_MPCAPABLEPASSIVEACK); + + newsk->sk_kern_sock = kern; + lock_sock(newsk); + __inet_accept(sock, newsock, newsk); set_bit(SOCK_CUSTOM_SOCKOPT, &newsock->flags); + msk = mptcp_sk(newsk); msk->in_accept_queue = 0; /* set ssk->sk_socket of accept()ed flows to mptcp socket. @@ -3881,6 +3846,21 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock, if (unlikely(list_is_singular(&msk->conn_list))) inet_sk_state_store(newsk, TCP_CLOSE); } + } else { + MPTCP_INC_STATS(sock_net(ssk), + MPTCP_MIB_MPCAPABLEPASSIVEFALLBACK); +tcpfallback: + newsk->sk_kern_sock = kern; + lock_sock(newsk); + __inet_accept(sock, newsock, newsk); + /* we are being invoked after accepting a non-mp-capable + * flow: sk is a tcp_sk, not an mptcp one. + * + * Hand the socket over to tcp so all further socket ops + * bypass mptcp. + */ + WRITE_ONCE(newsock->sk->sk_socket->ops, + mptcp_fallback_tcp_ops(newsock->sk)); } release_sock(newsk);