Received-SPF: pass (google.com: domain of linux-kernel+bounces-49775-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
From: "Matthieu Baerts (NGI0)" <matttbe@kernel.org>
Date: Fri, 02 Feb 2024 12:40:08 +0100
Subject: [PATCH net-next 2/5] mptcp: annotate lockless access for the tx
 path
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
Message-Id: <20240202-upstream-net-next-20240202-mptcp-annotate-lockless-access-v1-2-031d6680afdc@kernel.org>
References: <20240202-upstream-net-next-20240202-mptcp-annotate-lockless-access-v1-0-031d6680afdc@kernel.org>
In-Reply-To: <20240202-upstream-net-next-20240202-mptcp-annotate-lockless-access-v1-0-031d6680afdc@kernel.org>
To: mptcp@lists.linux.dev, Mat Martineau <martineau@kernel.org>, 
 Geliang Tang <geliang.tang@linux.dev>, 
 "David S. Miller" <davem@davemloft.net>, Eric Dumazet <edumazet@google.com>, 
 Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org, linux-kernel@vger.kernel.org, 
 "Matthieu Baerts (NGI0)" <matttbe@kernel.org>

From: Paolo Abeni <pabeni@redhat.com>

The mptcp-level TX path info (write_seq, bytes_sent, snd_nxt) are under
the msk socket lock protection, and are accessed lockless in a few spots.

Always mark the write operations with WRITE_ONCE, read operations
outside the lock with READ_ONCE and drop the annotation for read
under such lock.

To simplify the annotations move mptcp_pending_data_fin_ack() from
__mptcp_data_acked() to __mptcp_clean_una(), under the msk socket
lock, where such call would belong.

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
 net/mptcp/options.c  |  2 +-
 net/mptcp/protocol.c | 15 +++++++--------
 net/mptcp/protocol.h |  2 +-
 3 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/net/mptcp/options.c b/net/mptcp/options.c
index 5e2b130d8680..9b31a0a06265 100644
--- a/net/mptcp/options.c
+++ b/net/mptcp/options.c
@@ -1060,7 +1060,7 @@ static void ack_update_msk(struct mptcp_sock *msk,
 		msk->wnd_end = new_wnd_end;
 
 	/* this assumes mptcp_incoming_options() is invoked after tcp_ack() */
-	if (after64(msk->wnd_end, READ_ONCE(msk->snd_nxt)))
+	if (after64(msk->wnd_end, snd_nxt))
 		__mptcp_check_push(sk, ssk);
 
 	if (after64(new_snd_una, old_snd_una)) {
diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index 874f019c5093..b1c24ac3630f 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -1033,13 +1033,15 @@ static void __mptcp_clean_una(struct sock *sk)
 		msk->recovery = false;
 
 out:
-	if (snd_una == READ_ONCE(msk->snd_nxt) &&
-	    snd_una == READ_ONCE(msk->write_seq)) {
+	if (snd_una == msk->snd_nxt && snd_una == msk->write_seq) {
 		if (mptcp_rtx_timer_pending(sk) && !mptcp_data_fin_enabled(msk))
 			mptcp_stop_rtx_timer(sk);
 	} else {
 		mptcp_reset_rtx_timer(sk);
 	}
+
+	if (mptcp_pending_data_fin_ack(sk))
+		mptcp_schedule_work(sk);
 }
 
 static void __mptcp_clean_una_wakeup(struct sock *sk)
@@ -1499,7 +1501,7 @@ static void mptcp_update_post_push(struct mptcp_sock *msk,
 	 */
 	if (likely(after64(snd_nxt_new, msk->snd_nxt))) {
 		msk->bytes_sent += snd_nxt_new - msk->snd_nxt;
-		msk->snd_nxt = snd_nxt_new;
+		WRITE_ONCE(msk->snd_nxt, snd_nxt_new);
 	}
 }
 
@@ -3200,8 +3202,8 @@ struct sock *mptcp_sk_clone_init(const struct sock *sk,
 	if (mp_opt->suboptions & OPTION_MPTCP_CSUMREQD)
 		WRITE_ONCE(msk->csum_enabled, true);
 
-	msk->write_seq = subflow_req->idsn + 1;
-	msk->snd_nxt = msk->write_seq;
+	WRITE_ONCE(msk->write_seq, subflow_req->idsn + 1);
+	WRITE_ONCE(msk->snd_nxt, msk->write_seq);
 	msk->snd_una = msk->write_seq;
 	msk->wnd_end = msk->snd_nxt + req->rsk_rcv_wnd;
 	msk->setsockopt_seq = mptcp_sk(sk)->setsockopt_seq;
@@ -3303,9 +3305,6 @@ void __mptcp_data_acked(struct sock *sk)
 		__mptcp_clean_una(sk);
 	else
 		__set_bit(MPTCP_CLEAN_UNA, &mptcp_sk(sk)->cb_flags);
-
-	if (mptcp_pending_data_fin_ack(sk))
-		mptcp_schedule_work(sk);
 }
 
 void __mptcp_check_push(struct sock *sk, struct sock *ssk)
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 13b8cf8ec704..421dede93e2b 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -402,7 +402,7 @@ static inline struct mptcp_data_frag *mptcp_rtx_head(struct sock *sk)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
 
-	if (msk->snd_una == READ_ONCE(msk->snd_nxt))
+	if (msk->snd_una == msk->snd_nxt)
 		return NULL;
 
 	return list_first_entry_or_null(&msk->rtx_queue, struct mptcp_data_frag, list);

-- 
2.43.0