[PATCH net 0/3] mptcp: fix races on accept()

[PATCH net 0/3] mptcp: fix races on accept()

[Paolo Abeni]

This series includes some fixes for accept() races which may cause inconsistent
MPTCP socket status and oops. Please see the individual patches for the
technical details.

Florian Westphal (1):
  mptcp: handle mptcp listener destruction via rcu

Paolo Abeni (2):
  mptcp: avoid flipping mp_capable field in syn_recv_sock()
  mptcp: drop req socket remote_key* fields

 net/mptcp/protocol.c | 11 ++++++--
 net/mptcp/protocol.h |  8 +++---
 net/mptcp/subflow.c  | 66 +++++++++++++++++++++++++++-----------------
 3 files changed, 52 insertions(+), 33 deletions(-)

-- 
2.21.1

[PATCH net 1/3] mptcp: handle mptcp listener destruction via rcu

[Paolo Abeni]

From: Florian Westphal <fw@strlen.de>

Following splat can occur during self test:

 BUG: KASAN: use-after-free in subflow_data_ready+0x156/0x160
 Read of size 8 at addr ffff888100c35c28 by task mptcp_connect/4808

  subflow_data_ready+0x156/0x160
  tcp_child_process+0x6a3/0xb30
  tcp_v4_rcv+0x2231/0x3730
  ip_protocol_deliver_rcu+0x5c/0x860
  ip_local_deliver_finish+0x220/0x360
  ip_local_deliver+0x1c8/0x4e0
  ip_rcv_finish+0x1da/0x2f0
  ip_rcv+0xd0/0x3c0
  __netif_receive_skb_one_core+0xf5/0x160
  __netif_receive_skb+0x27/0x1c0
  process_backlog+0x21e/0x780
  net_rx_action+0x35f/0xe90
  do_softirq+0x4c/0x50
  [..]

This occurs when accessing subflow_ctx->conn.

Problem is that tcp_child_process() calls listen sockets'
sk_data_ready() notification, but it doesn't hold the listener
lock.  Another cpu calling close() on the listener will then cause
transition of refcount to 0.

Fixes: 58b09919626bf ("mptcp: create msk early")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
 net/mptcp/protocol.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index b2d2193453bc..d275c1e827fe 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -1391,6 +1391,7 @@ struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
 		msk->ack_seq = ack_seq;
 	}
 
+	sock_reset_flag(nsk, SOCK_RCU_FREE);
 	/* will be fully established after successful MPC subflow creation */
 	inet_sk_state_store(nsk, TCP_SYN_RECV);
 	bh_unlock_sock(nsk);
@@ -1792,6 +1793,8 @@ static int mptcp_listen(struct socket *sock, int backlog)
 		goto unlock;
 	}
 
+	sock_set_flag(sock->sk, SOCK_RCU_FREE);
+
 	err = ssock->ops->listen(ssock, backlog);
 	inet_sk_state_store(sock->sk, inet_sk_state_load(ssock->sk));
 	if (!err)
-- 
2.21.1

[PATCH net 2/3] mptcp: avoid flipping mp_capable field in syn_recv_sock()

[Paolo Abeni]

If multiple CPUs races on the same req_sock in syn_recv_sock(),
flipping such field can cause inconsistent child socket status.

When racing, the CPU losing the req ownership may still change
the mptcp request socket mp_capable flag while the CPU owning
the request is cloning the socket, leaving the child socket with
'is_mptcp' set but no 'mp_capable' flag.

Such socket will stay with 'conn' field cleared, heading to oops
in later mptcp callback.

Address the issue tracking the fallback status in a local variable.

Fixes: 58b09919626b ("mptcp: create msk early")
Co-developed-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
 net/mptcp/subflow.c | 46 +++++++++++++++++++++++++++++----------------
 1 file changed, 30 insertions(+), 16 deletions(-)

diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
index 3a94f859347a..10090ca3d3e0 100644
--- a/net/mptcp/subflow.c
+++ b/net/mptcp/subflow.c
@@ -376,6 +376,17 @@ static void mptcp_force_close(struct sock *sk)
 	sk_common_release(sk);
 }
 
+static void subflow_ulp_fallback(struct sock *sk,
+				 struct mptcp_subflow_context *old_ctx)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+
+	mptcp_subflow_tcp_fallback(sk, old_ctx);
+	icsk->icsk_ulp_ops = NULL;
+	rcu_assign_pointer(icsk->icsk_ulp_data, NULL);
+	tcp_sk(sk)->is_mptcp = 0;
+}
+
 static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 					  struct sk_buff *skb,
 					  struct request_sock *req,
@@ -388,6 +399,7 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 	struct tcp_options_received opt_rx;
 	bool fallback_is_fatal = false;
 	struct sock *new_msk = NULL;
+	bool fallback = false;
 	struct sock *child;
 
 	pr_debug("listener=%p, req=%p, conn=%p", listener, req, listener->conn);
@@ -412,14 +424,14 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 			subflow_req->remote_key = opt_rx.mptcp.sndr_key;
 			subflow_req->remote_key_valid = 1;
 		} else {
-			subflow_req->mp_capable = 0;
+			fallback = true;
 			goto create_child;
 		}
 
 create_msk:
 		new_msk = mptcp_sk_clone(listener->conn, req);
 		if (!new_msk)
-			subflow_req->mp_capable = 0;
+			fallback = true;
 	} else if (subflow_req->mp_join) {
 		fallback_is_fatal = true;
 		opt_rx.mptcp.mp_join = 0;
@@ -438,12 +450,18 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 	if (child && *own_req) {
 		struct mptcp_subflow_context *ctx = mptcp_subflow_ctx(child);
 
-		/* we have null ctx on TCP fallback, which is fatal on
-		 * MPJ handshake
+		/* we need to fallback on ctx allocation failure and on pre-reqs
+		 * checking above. In the latter scenario we additionally need
+		 * to reset the context to non MPTCP status.
 		 */
-		if (!ctx) {
+		if (!ctx || fallback) {
 			if (fallback_is_fatal)
 				goto close_child;
+
+			if (ctx) {
+				subflow_ulp_fallback(child, ctx);
+				kfree_rcu(ctx, rcu);
+			}
 			goto out;
 		}
 
@@ -474,6 +492,13 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 	/* dispose of the left over mptcp master, if any */
 	if (unlikely(new_msk))
 		mptcp_force_close(new_msk);
+
+	/* check for expected invariant - should never trigger, just help
+	 * catching eariler subtle bugs
+	 */
+	WARN_ON_ONCE(*own_req && child && tcp_sk(child)->is_mptcp &&
+		     (!mptcp_subflow_ctx(child) ||
+		      !mptcp_subflow_ctx(child)->conn));
 	return child;
 
 close_child:
@@ -1094,17 +1119,6 @@ static void subflow_ulp_release(struct sock *sk)
 	kfree_rcu(ctx, rcu);
 }
 
-static void subflow_ulp_fallback(struct sock *sk,
-				 struct mptcp_subflow_context *old_ctx)
-{
-	struct inet_connection_sock *icsk = inet_csk(sk);
-
-	mptcp_subflow_tcp_fallback(sk, old_ctx);
-	icsk->icsk_ulp_ops = NULL;
-	rcu_assign_pointer(icsk->icsk_ulp_data, NULL);
-	tcp_sk(sk)->is_mptcp = 0;
-}
-
 static void subflow_ulp_clone(const struct request_sock *req,
 			      struct sock *newsk,
 			      const gfp_t priority)
-- 
2.21.1

[PATCH net 3/3] mptcp: drop req socket remote_key* fields

[Paolo Abeni]

We don't need them, as we can use the current ingress opt
data instead. Setting them in syn_recv_sock() may causes
inconsistent mptcp socket status, as per previous commit.

Fixes: cc7972ea1932 ("mptcp: parse and emit MP_CAPABLE option according to v1 spec")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
 net/mptcp/protocol.c |  8 +++++---
 net/mptcp/protocol.h |  8 ++++----
 net/mptcp/subflow.c  | 20 ++++++++++----------
 3 files changed, 19 insertions(+), 17 deletions(-)

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index d275c1e827fe..58ad03fc1bbc 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -1345,7 +1345,9 @@ static struct ipv6_pinfo *mptcp_inet6_sk(const struct sock *sk)
 }
 #endif
 
-struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
+struct sock *mptcp_sk_clone(const struct sock *sk,
+			    const struct tcp_options_received *opt_rx,
+			    struct request_sock *req)
 {
 	struct mptcp_subflow_request_sock *subflow_req = mptcp_subflow_rsk(req);
 	struct sock *nsk = sk_clone_lock(sk, GFP_ATOMIC);
@@ -1383,9 +1385,9 @@ struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
 
 	msk->write_seq = subflow_req->idsn + 1;
 	atomic64_set(&msk->snd_una, msk->write_seq);
-	if (subflow_req->remote_key_valid) {
+	if (opt_rx->mptcp.mp_capable) {
 		msk->can_ack = true;
-		msk->remote_key = subflow_req->remote_key;
+		msk->remote_key = opt_rx->mptcp.sndr_key;
 		mptcp_crypto_key_sha(msk->remote_key, NULL, &ack_seq);
 		ack_seq++;
 		msk->ack_seq = ack_seq;
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 67448002a2d7..a2b3048037d0 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -206,12 +206,10 @@ struct mptcp_subflow_request_sock {
 	struct	tcp_request_sock sk;
 	u16	mp_capable : 1,
 		mp_join : 1,
-		backup : 1,
-		remote_key_valid : 1;
+		backup : 1;
 	u8	local_id;
 	u8	remote_id;
 	u64	local_key;
-	u64	remote_key;
 	u64	idsn;
 	u32	token;
 	u32	ssn_offset;
@@ -332,7 +330,9 @@ void mptcp_proto_init(void);
 int mptcp_proto_v6_init(void);
 #endif
 
-struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req);
+struct sock *mptcp_sk_clone(const struct sock *sk,
+			    const struct tcp_options_received *opt_rx,
+			    struct request_sock *req);
 void mptcp_get_options(const struct sk_buff *skb,
 		       struct tcp_options_received *opt_rx);
 
diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
index 10090ca3d3e0..87c094702d63 100644
--- a/net/mptcp/subflow.c
+++ b/net/mptcp/subflow.c
@@ -133,7 +133,6 @@ static void subflow_init_req(struct request_sock *req,
 
 	subflow_req->mp_capable = 0;
 	subflow_req->mp_join = 0;
-	subflow_req->remote_key_valid = 0;
 
 #ifdef CONFIG_TCP_MD5SIG
 	/* no MPTCP if MD5SIG is enabled on this socket or we may run out of
@@ -404,6 +403,7 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 
 	pr_debug("listener=%p, req=%p, conn=%p", listener, req, listener->conn);
 
+	opt_rx.mptcp.mp_capable = 0;
 	if (tcp_rsk(req)->is_mptcp == 0)
 		goto create_child;
 
@@ -418,18 +418,14 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 			goto create_msk;
 		}
 
-		opt_rx.mptcp.mp_capable = 0;
 		mptcp_get_options(skb, &opt_rx);
-		if (opt_rx.mptcp.mp_capable) {
-			subflow_req->remote_key = opt_rx.mptcp.sndr_key;
-			subflow_req->remote_key_valid = 1;
-		} else {
+		if (!opt_rx.mptcp.mp_capable) {
 			fallback = true;
 			goto create_child;
 		}
 
 create_msk:
-		new_msk = mptcp_sk_clone(listener->conn, req);
+		new_msk = mptcp_sk_clone(listener->conn, &opt_rx, req);
 		if (!new_msk)
 			fallback = true;
 	} else if (subflow_req->mp_join) {
@@ -473,6 +469,13 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 			mptcp_pm_new_connection(mptcp_sk(new_msk), 1);
 			ctx->conn = new_msk;
 			new_msk = NULL;
+
+			/* with OoO packets we can reach here without ingress
+			 * mpc option
+			 */
+			ctx->remote_key = opt_rx.mptcp.sndr_key;
+			ctx->fully_established = opt_rx.mptcp.mp_capable;
+			ctx->can_ack = opt_rx.mptcp.mp_capable;
 		} else if (ctx->mp_join) {
 			struct mptcp_sock *owner;
 
@@ -1152,9 +1155,6 @@ static void subflow_ulp_clone(const struct request_sock *req,
 		 * is fully established only after we receive the remote key
 		 */
 		new_ctx->mp_capable = 1;
-		new_ctx->fully_established = subflow_req->remote_key_valid;
-		new_ctx->can_ack = subflow_req->remote_key_valid;
-		new_ctx->remote_key = subflow_req->remote_key;
 		new_ctx->local_key = subflow_req->local_key;
 		new_ctx->token = subflow_req->token;
 		new_ctx->ssn_offset = subflow_req->ssn_offset;
-- 
2.21.1

Re: [PATCH net 0/3] mptcp: fix races on accept()

[David Miller]

From: Paolo Abeni <pabeni@redhat.com>
Date: Mon, 20 Apr 2020 16:25:03 +0200

> This series includes some fixes for accept() races which may cause inconsistent
> MPTCP socket status and oops. Please see the individual patches for the
> technical details.

Series applied, thanks.

It seems like patch #3 might be relevant for v5.6 -stable, what's the
story here?

Re: [PATCH net 3/3] mptcp: drop req socket remote_key* fields

[Mat Martineau]

On Mon, 20 Apr 2020, Paolo Abeni wrote:

> We don't need them, as we can use the current ingress opt
> data instead. Setting them in syn_recv_sock() may causes
> inconsistent mptcp socket status, as per previous commit.
>
> Fixes: cc7972ea1932 ("mptcp: parse and emit MP_CAPABLE option according to v1 spec")
> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
> ---
> net/mptcp/protocol.c |  8 +++++---
> net/mptcp/protocol.h |  8 ++++----
> net/mptcp/subflow.c  | 20 ++++++++++----------
> 3 files changed, 19 insertions(+), 17 deletions(-)
>
> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> index d275c1e827fe..58ad03fc1bbc 100644
> --- a/net/mptcp/protocol.c
> +++ b/net/mptcp/protocol.c
> @@ -1345,7 +1345,9 @@ static struct ipv6_pinfo *mptcp_inet6_sk(const struct sock *sk)
> }
> #endif
>
> -struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
> +struct sock *mptcp_sk_clone(const struct sock *sk,
> +			    const struct tcp_options_received *opt_rx,
> +			    struct request_sock *req)
> {
> 	struct mptcp_subflow_request_sock *subflow_req = mptcp_subflow_rsk(req);
> 	struct sock *nsk = sk_clone_lock(sk, GFP_ATOMIC);
> @@ -1383,9 +1385,9 @@ struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
>
> 	msk->write_seq = subflow_req->idsn + 1;
> 	atomic64_set(&msk->snd_una, msk->write_seq);
> -	if (subflow_req->remote_key_valid) {
> +	if (opt_rx->mptcp.mp_capable) {
> 		msk->can_ack = true;
> -		msk->remote_key = subflow_req->remote_key;
> +		msk->remote_key = opt_rx->mptcp.sndr_key;
> 		mptcp_crypto_key_sha(msk->remote_key, NULL, &ack_seq);
> 		ack_seq++;
> 		msk->ack_seq = ack_seq;
> diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> index 67448002a2d7..a2b3048037d0 100644
> --- a/net/mptcp/protocol.h
> +++ b/net/mptcp/protocol.h
> @@ -206,12 +206,10 @@ struct mptcp_subflow_request_sock {
> 	struct	tcp_request_sock sk;
> 	u16	mp_capable : 1,
> 		mp_join : 1,
> -		backup : 1,
> -		remote_key_valid : 1;
> +		backup : 1;
> 	u8	local_id;
> 	u8	remote_id;
> 	u64	local_key;
> -	u64	remote_key;
> 	u64	idsn;
> 	u32	token;
> 	u32	ssn_offset;
> @@ -332,7 +330,9 @@ void mptcp_proto_init(void);
> int mptcp_proto_v6_init(void);
> #endif
>
> -struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req);
> +struct sock *mptcp_sk_clone(const struct sock *sk,
> +			    const struct tcp_options_received *opt_rx,
> +			    struct request_sock *req);
> void mptcp_get_options(const struct sk_buff *skb,
> 		       struct tcp_options_received *opt_rx);
>
> diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
> index 10090ca3d3e0..87c094702d63 100644
> --- a/net/mptcp/subflow.c
> +++ b/net/mptcp/subflow.c
> @@ -133,7 +133,6 @@ static void subflow_init_req(struct request_sock *req,
>
> 	subflow_req->mp_capable = 0;
> 	subflow_req->mp_join = 0;
> -	subflow_req->remote_key_valid = 0;
>
> #ifdef CONFIG_TCP_MD5SIG
> 	/* no MPTCP if MD5SIG is enabled on this socket or we may run out of
> @@ -404,6 +403,7 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
>
> 	pr_debug("listener=%p, req=%p, conn=%p", listener, req, listener->conn);
>
> +	opt_rx.mptcp.mp_capable = 0;
> 	if (tcp_rsk(req)->is_mptcp == 0)
> 		goto create_child;
>
> @@ -418,18 +418,14 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> 			goto create_msk;
> 		}
>
> -		opt_rx.mptcp.mp_capable = 0;
> 		mptcp_get_options(skb, &opt_rx);
> -		if (opt_rx.mptcp.mp_capable) {
> -			subflow_req->remote_key = opt_rx.mptcp.sndr_key;
> -			subflow_req->remote_key_valid = 1;
> -		} else {
> +		if (!opt_rx.mptcp.mp_capable) {
> 			fallback = true;
> 			goto create_child;
> 		}
>
> create_msk:
> -		new_msk = mptcp_sk_clone(listener->conn, req);
> +		new_msk = mptcp_sk_clone(listener->conn, &opt_rx, req);
> 		if (!new_msk)
> 			fallback = true;
> 	} else if (subflow_req->mp_join) {
> @@ -473,6 +469,13 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> 			mptcp_pm_new_connection(mptcp_sk(new_msk), 1);
> 			ctx->conn = new_msk;
> 			new_msk = NULL;
> +
> +			/* with OoO packets we can reach here without ingress
> +			 * mpc option
> +			 */
> +			ctx->remote_key = opt_rx.mptcp.sndr_key;
> +			ctx->fully_established = opt_rx.mptcp.mp_capable;
> +			ctx->can_ack = opt_rx.mptcp.mp_capable;

If this code is reached without an incoming mpc option, it looks like 
ctx->remote_key gets junk off the stack. Maybe this instead?

+			if (opt_rx.mptcp.mp_capable) {
+				ctx->remote_key = opt_rx.mptcp.sndr_key;
+				ctx->fully_established = 1;
+				ctx->can_ack = 1;
+			}


Mat

> 		} else if (ctx->mp_join) {
> 			struct mptcp_sock *owner;
>
> @@ -1152,9 +1155,6 @@ static void subflow_ulp_clone(const struct request_sock *req,
> 		 * is fully established only after we receive the remote key
> 		 */
> 		new_ctx->mp_capable = 1;
> -		new_ctx->fully_established = subflow_req->remote_key_valid;
> -		new_ctx->can_ack = subflow_req->remote_key_valid;
> -		new_ctx->remote_key = subflow_req->remote_key;
> 		new_ctx->local_key = subflow_req->local_key;
> 		new_ctx->token = subflow_req->token;
> 		new_ctx->ssn_offset = subflow_req->ssn_offset;
> -- 
> 2.21.1
>
>

--
Mat Martineau
Intel

Re: [PATCH net 3/3] mptcp: drop req socket remote_key* fields

[Paolo Abeni]

On Mon, 2020-04-20 at 13:05 -0700, Mat Martineau wrote:
> On Mon, 20 Apr 2020, Paolo Abeni wrote:
> 
> > We don't need them, as we can use the current ingress opt
> > data instead. Setting them in syn_recv_sock() may causes
> > inconsistent mptcp socket status, as per previous commit.
> > 
> > Fixes: cc7972ea1932 ("mptcp: parse and emit MP_CAPABLE option according to v1 spec")
> > Signed-off-by: Paolo Abeni <pabeni@redhat.com>
> > ---
> > net/mptcp/protocol.c |  8 +++++---
> > net/mptcp/protocol.h |  8 ++++----
> > net/mptcp/subflow.c  | 20 ++++++++++----------
> > 3 files changed, 19 insertions(+), 17 deletions(-)
> > 
> > diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> > index d275c1e827fe..58ad03fc1bbc 100644
> > --- a/net/mptcp/protocol.c
> > +++ b/net/mptcp/protocol.c
> > @@ -1345,7 +1345,9 @@ static struct ipv6_pinfo *mptcp_inet6_sk(const struct sock *sk)
> > }
> > #endif
> > 
> > -struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
> > +struct sock *mptcp_sk_clone(const struct sock *sk,
> > +			    const struct tcp_options_received *opt_rx,
> > +			    struct request_sock *req)
> > {
> > 	struct mptcp_subflow_request_sock *subflow_req = mptcp_subflow_rsk(req);
> > 	struct sock *nsk = sk_clone_lock(sk, GFP_ATOMIC);
> > @@ -1383,9 +1385,9 @@ struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req)
> > 
> > 	msk->write_seq = subflow_req->idsn + 1;
> > 	atomic64_set(&msk->snd_una, msk->write_seq);
> > -	if (subflow_req->remote_key_valid) {
> > +	if (opt_rx->mptcp.mp_capable) {
> > 		msk->can_ack = true;
> > -		msk->remote_key = subflow_req->remote_key;
> > +		msk->remote_key = opt_rx->mptcp.sndr_key;
> > 		mptcp_crypto_key_sha(msk->remote_key, NULL, &ack_seq);
> > 		ack_seq++;
> > 		msk->ack_seq = ack_seq;
> > diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> > index 67448002a2d7..a2b3048037d0 100644
> > --- a/net/mptcp/protocol.h
> > +++ b/net/mptcp/protocol.h
> > @@ -206,12 +206,10 @@ struct mptcp_subflow_request_sock {
> > 	struct	tcp_request_sock sk;
> > 	u16	mp_capable : 1,
> > 		mp_join : 1,
> > -		backup : 1,
> > -		remote_key_valid : 1;
> > +		backup : 1;
> > 	u8	local_id;
> > 	u8	remote_id;
> > 	u64	local_key;
> > -	u64	remote_key;
> > 	u64	idsn;
> > 	u32	token;
> > 	u32	ssn_offset;
> > @@ -332,7 +330,9 @@ void mptcp_proto_init(void);
> > int mptcp_proto_v6_init(void);
> > #endif
> > 
> > -struct sock *mptcp_sk_clone(const struct sock *sk, struct request_sock *req);
> > +struct sock *mptcp_sk_clone(const struct sock *sk,
> > +			    const struct tcp_options_received *opt_rx,
> > +			    struct request_sock *req);
> > void mptcp_get_options(const struct sk_buff *skb,
> > 		       struct tcp_options_received *opt_rx);
> > 
> > diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
> > index 10090ca3d3e0..87c094702d63 100644
> > --- a/net/mptcp/subflow.c
> > +++ b/net/mptcp/subflow.c
> > @@ -133,7 +133,6 @@ static void subflow_init_req(struct request_sock *req,
> > 
> > 	subflow_req->mp_capable = 0;
> > 	subflow_req->mp_join = 0;
> > -	subflow_req->remote_key_valid = 0;
> > 
> > #ifdef CONFIG_TCP_MD5SIG
> > 	/* no MPTCP if MD5SIG is enabled on this socket or we may run out of
> > @@ -404,6 +403,7 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > 
> > 	pr_debug("listener=%p, req=%p, conn=%p", listener, req, listener->conn);
> > 
> > +	opt_rx.mptcp.mp_capable = 0;
> > 	if (tcp_rsk(req)->is_mptcp == 0)
> > 		goto create_child;
> > 
> > @@ -418,18 +418,14 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > 			goto create_msk;
> > 		}
> > 
> > -		opt_rx.mptcp.mp_capable = 0;
> > 		mptcp_get_options(skb, &opt_rx);
> > -		if (opt_rx.mptcp.mp_capable) {
> > -			subflow_req->remote_key = opt_rx.mptcp.sndr_key;
> > -			subflow_req->remote_key_valid = 1;
> > -		} else {
> > +		if (!opt_rx.mptcp.mp_capable) {
> > 			fallback = true;
> > 			goto create_child;
> > 		}
> > 
> > create_msk:
> > -		new_msk = mptcp_sk_clone(listener->conn, req);
> > +		new_msk = mptcp_sk_clone(listener->conn, &opt_rx, req);
> > 		if (!new_msk)
> > 			fallback = true;
> > 	} else if (subflow_req->mp_join) {
> > @@ -473,6 +469,13 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
> > 			mptcp_pm_new_connection(mptcp_sk(new_msk), 1);
> > 			ctx->conn = new_msk;
> > 			new_msk = NULL;
> > +
> > +			/* with OoO packets we can reach here without ingress
> > +			 * mpc option
> > +			 */
> > +			ctx->remote_key = opt_rx.mptcp.sndr_key;
> > +			ctx->fully_established = opt_rx.mptcp.mp_capable;
> > +			ctx->can_ack = opt_rx.mptcp.mp_capable;
> 
> If this code is reached without an incoming mpc option, it looks like 
> ctx->remote_key gets junk off the stack. Maybe this instead?

The idea here is to avoid additional conditional. The remote_key will
be used only after 'can_ack' becomes true, and the '!can_ack' condition
here is unlikely. Overall this should be faster and safe.

Thanks for checking,

Paolo

Re: [PATCH net 0/3] mptcp: fix races on accept()

[Paolo Abeni]

On Mon, 2020-04-20 at 13:02 -0700, David Miller wrote:
> From: Paolo Abeni <pabeni@redhat.com>
> Date: Mon, 20 Apr 2020 16:25:03 +0200
> 
> > This series includes some fixes for accept() races which may cause inconsistent
> > MPTCP socket status and oops. Please see the individual patches for the
> > technical details.
> 
> Series applied, thanks.
> 
> It seems like patch #3 might be relevant for v5.6 -stable, what's the
> story here?

Yes, it addresses a race condition present since cc7972ea1932 ("mptcp:
parse and emit MP_CAPABLE option according to v1 spec"). I see now that
the changelog is probably a bit too vague, I'm sorry.

Cheers,

Paolo