From 7c85af8810448d8ef59331be51e482413b5f503d Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Thu, 24 Sep 2015 17:16:05 -0700
Subject: tcp: avoid reorders for TFO passive connections

We found that a TCP Fast Open passive connection was vulnerable
to reorders, as the exchange might look like

[1] C -> S S <FO ...> <request>
[2] S -> C S. ack request <options>
[3] S -> C . <answer>

packets [2] and [3] can be generated at almost the same time.

If C receives the 3rd packet before the 2nd, it will drop it as
the socket is in SYN_SENT state and expects a SYNACK.

S will have to retransmit the answer.

Current OOO avoidance in linux is defeated because SYNACK
packets are attached to the LISTEN socket, while DATA packets
are attached to the children. They might be sent by different cpus,
and different TX queues might be selected.

It turns out that for TFO, we created a child, which is a
full blown socket in TCP_SYN_RECV state, and we simply can attach
the SYNACK packet to this socket.

This means that at the time tcp_sendmsg() pushes DATA packet,
skb->ooo_okay will be set iff the SYNACK packet had been sent
and TX completed.

This removes the reorder source at the host level.

We also removed the export of tcp_try_fastopen(), as it is no
longer called from IPv6.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 net/ipv4/tcp_fastopen.c | 35 +++++++++++++++++++----------------
 1 file changed, 19 insertions(+), 16 deletions(-)

(limited to 'net/ipv4/tcp_fastopen.c')

diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
index f9c0fb84e435..db43c6286cf7 100644
--- a/net/ipv4/tcp_fastopen.c
+++ b/net/ipv4/tcp_fastopen.c
@@ -124,10 +124,10 @@ static bool tcp_fastopen_cookie_gen(struct request_sock *req,
 	return false;
 }
 
-static bool tcp_fastopen_create_child(struct sock *sk,
-				      struct sk_buff *skb,
-				      struct dst_entry *dst,
-				      struct request_sock *req)
+static struct sock *tcp_fastopen_create_child(struct sock *sk,
+					      struct sk_buff *skb,
+					      struct dst_entry *dst,
+					      struct request_sock *req)
 {
 	struct tcp_sock *tp;
 	struct request_sock_queue *queue = &inet_csk(sk)->icsk_accept_queue;
@@ -140,7 +140,7 @@ static bool tcp_fastopen_create_child(struct sock *sk,
 
 	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
 	if (!child)
-		return false;
+		return NULL;
 
 	spin_lock(&queue->fastopenq->lock);
 	queue->fastopenq->qlen++;
@@ -216,9 +216,11 @@ static bool tcp_fastopen_create_child(struct sock *sk,
 	tcp_rsk(req)->rcv_nxt = tp->rcv_nxt = end_seq;
 	sk->sk_data_ready(sk);
 	bh_unlock_sock(child);
-	sock_put(child);
+	/* Note: sock_put(child) will be done by tcp_conn_request()
+	 * after SYNACK packet is sent.
+	 */
 	WARN_ON(!req->sk);
-	return true;
+	return child;
 }
 
 static bool tcp_fastopen_queue_check(struct sock *sk)
@@ -261,13 +263,14 @@ static bool tcp_fastopen_queue_check(struct sock *sk)
  * may be updated and return the client in the SYN-ACK later. E.g., Fast Open
  * cookie request (foc->len == 0).
  */
-bool tcp_try_fastopen(struct sock *sk, struct sk_buff *skb,
-		      struct request_sock *req,
-		      struct tcp_fastopen_cookie *foc,
-		      struct dst_entry *dst)
+struct sock *tcp_try_fastopen(struct sock *sk, struct sk_buff *skb,
+			      struct request_sock *req,
+			      struct tcp_fastopen_cookie *foc,
+			      struct dst_entry *dst)
 {
 	struct tcp_fastopen_cookie valid_foc = { .len = -1 };
 	bool syn_data = TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq + 1;
+	struct sock *child;
 
 	if (foc->len == 0) /* Client requests a cookie */
 		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENCOOKIEREQD);
@@ -276,7 +279,7 @@ bool tcp_try_fastopen(struct sock *sk, struct sk_buff *skb,
 	      (syn_data || foc->len >= 0) &&
 	      tcp_fastopen_queue_check(sk))) {
 		foc->len = -1;
-		return false;
+		return NULL;
 	}
 
 	if (syn_data && (sysctl_tcp_fastopen & TFO_SERVER_COOKIE_NOT_REQD))
@@ -296,11 +299,12 @@ bool tcp_try_fastopen(struct sock *sk, struct sk_buff *skb,
 		 * data in SYN_RECV state.
 		 */
 fastopen:
-		if (tcp_fastopen_create_child(sk, skb, dst, req)) {
+		child = tcp_fastopen_create_child(sk, skb, dst, req);
+		if (child) {
 			foc->len = -1;
 			NET_INC_STATS_BH(sock_net(sk),
 					 LINUX_MIB_TCPFASTOPENPASSIVE);
-			return true;
+			return child;
 		}
 		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENPASSIVEFAIL);
 	} else if (foc->len > 0) /* Client presents an invalid cookie */
@@ -308,6 +312,5 @@ fastopen:
 
 	valid_foc.exp = foc->exp;
 	*foc = valid_foc;
-	return false;
+	return NULL;
 }
-EXPORT_SYMBOL(tcp_try_fastopen);
-- 
cgit v1.2.3


From 0536fcc039a8926ec12ec587f41a83f7acafeb82 Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Tue, 29 Sep 2015 07:42:52 -0700
Subject: tcp: prepare fastopen code for upcoming listener changes

While auditing TCP stack for upcoming 'lockless' listener changes,
I found I had to change fastopen_init_queue() to properly init the object
before publishing it.

Otherwise an other cpu could try to lock the spinlock before it gets
properly initialized.

Instead of adding appropriate barriers, just remove dynamic memory
allocations :
- Structure is 28 bytes on 64bit arches. Using additional 8 bytes
  for holding a pointer seems overkill.
- Two listeners can share same cache line and performance would suffer.

If we really want to save few bytes, we would instead dynamically allocate
whole struct request_sock_queue in the future.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/linux/tcp.h             | 22 ++++------------------
 include/net/request_sock.h      |  7 ++-----
 net/core/request_sock.c         |  9 ++++++++-
 net/ipv4/af_inet.c              | 10 +++-------
 net/ipv4/inet_connection_sock.c | 17 ++++++++---------
 net/ipv4/tcp.c                  | 14 ++------------
 net/ipv4/tcp_fastopen.c         | 10 +++++-----
 net/ipv4/tcp_ipv4.c             |  2 +-
 net/ipv6/tcp_ipv6.c             |  4 ++--
 9 files changed, 35 insertions(+), 60 deletions(-)

(limited to 'net/ipv4/tcp_fastopen.c')

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index fcb573be75d9..e442e6e9a365 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -382,25 +382,11 @@ static inline bool tcp_passive_fastopen(const struct sock *sk)
 		tcp_sk(sk)->fastopen_rsk != NULL);
 }
 
-extern void tcp_sock_destruct(struct sock *sk);
-
-static inline int fastopen_init_queue(struct sock *sk, int backlog)
+static inline void fastopen_queue_tune(struct sock *sk, int backlog)
 {
-	struct request_sock_queue *queue =
-	    &inet_csk(sk)->icsk_accept_queue;
-
-	if (queue->fastopenq == NULL) {
-		queue->fastopenq = kzalloc(
-		    sizeof(struct fastopen_queue),
-		    sk->sk_allocation);
-		if (queue->fastopenq == NULL)
-			return -ENOMEM;
-
-		sk->sk_destruct = tcp_sock_destruct;
-		spin_lock_init(&queue->fastopenq->lock);
-	}
-	queue->fastopenq->max_qlen = backlog;
-	return 0;
+	struct request_sock_queue *queue = &inet_csk(sk)->icsk_accept_queue;
+
+	queue->fastopenq.max_qlen = backlog;
 }
 
 static inline void tcp_saved_syn_free(struct tcp_sock *tp)
diff --git a/include/net/request_sock.h b/include/net/request_sock.h
index c146b5284786..d2544de329bd 100644
--- a/include/net/request_sock.h
+++ b/include/net/request_sock.h
@@ -180,11 +180,8 @@ struct request_sock_queue {
 	struct request_sock	*rskq_accept_tail;
 	u8			rskq_defer_accept;
 	struct listen_sock	*listen_opt;
-	struct fastopen_queue	*fastopenq; /* This is non-NULL iff TFO has been
-					     * enabled on this listener. Check
-					     * max_qlen != 0 in fastopen_queue
-					     * to determine if TFO is enabled
-					     * right at this moment.
+	struct fastopen_queue	fastopenq;  /* Check max_qlen != 0 to determine
+					     * if TFO is enabled.
 					     */
 
 	/* temporary alignment, our goal is to get rid of this lock */
diff --git a/net/core/request_sock.c b/net/core/request_sock.c
index b42f0e26f89e..e22cfa4ed25f 100644
--- a/net/core/request_sock.c
+++ b/net/core/request_sock.c
@@ -59,6 +59,13 @@ int reqsk_queue_alloc(struct request_sock_queue *queue,
 
 	get_random_bytes(&lopt->hash_rnd, sizeof(lopt->hash_rnd));
 	spin_lock_init(&queue->syn_wait_lock);
+
+	spin_lock_init(&queue->fastopenq.lock);
+	queue->fastopenq.rskq_rst_head = NULL;
+	queue->fastopenq.rskq_rst_tail = NULL;
+	queue->fastopenq.qlen = 0;
+	queue->fastopenq.max_qlen = 0;
+
 	queue->rskq_accept_head = NULL;
 	lopt->nr_table_entries = nr_table_entries;
 	lopt->max_qlen_log = ilog2(nr_table_entries);
@@ -174,7 +181,7 @@ void reqsk_fastopen_remove(struct sock *sk, struct request_sock *req,
 	struct sock *lsk = req->rsk_listener;
 	struct fastopen_queue *fastopenq;
 
-	fastopenq = inet_csk(lsk)->icsk_accept_queue.fastopenq;
+	fastopenq = &inet_csk(lsk)->icsk_accept_queue.fastopenq;
 
 	tcp_sk(sk)->fastopen_rsk = NULL;
 	spin_lock_bh(&fastopenq->lock);
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 8a556643b874..3af85eecbe11 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -219,17 +219,13 @@ int inet_listen(struct socket *sock, int backlog)
 		 * shutdown() (rather than close()).
 		 */
 		if ((sysctl_tcp_fastopen & TFO_SERVER_ENABLE) != 0 &&
-		    !inet_csk(sk)->icsk_accept_queue.fastopenq) {
+		    !inet_csk(sk)->icsk_accept_queue.fastopenq.max_qlen) {
 			if ((sysctl_tcp_fastopen & TFO_SERVER_WO_SOCKOPT1) != 0)
-				err = fastopen_init_queue(sk, backlog);
+				fastopen_queue_tune(sk, backlog);
 			else if ((sysctl_tcp_fastopen &
 				  TFO_SERVER_WO_SOCKOPT2) != 0)
-				err = fastopen_init_queue(sk,
+				fastopen_queue_tune(sk,
 				    ((uint)sysctl_tcp_fastopen) >> 16);
-			else
-				err = 0;
-			if (err)
-				goto out;
 
 			tcp_fastopen_init_key_once(true);
 		}
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 694a5e8f4f9f..e1527882a578 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -335,9 +335,8 @@ struct sock *inet_csk_accept(struct sock *sk, int flags, int *err)
 
 	sk_acceptq_removed(sk);
 	if (sk->sk_protocol == IPPROTO_TCP &&
-	    tcp_rsk(req)->tfo_listener &&
-	    queue->fastopenq) {
-		spin_lock_bh(&queue->fastopenq->lock);
+	    tcp_rsk(req)->tfo_listener) {
+		spin_lock_bh(&queue->fastopenq.lock);
 		if (tcp_rsk(req)->tfo_listener) {
 			/* We are still waiting for the final ACK from 3WHS
 			 * so can't free req now. Instead, we set req->sk to
@@ -348,7 +347,7 @@ struct sock *inet_csk_accept(struct sock *sk, int flags, int *err)
 			req->sk = NULL;
 			req = NULL;
 		}
-		spin_unlock_bh(&queue->fastopenq->lock);
+		spin_unlock_bh(&queue->fastopenq.lock);
 	}
 out:
 	release_sock(sk);
@@ -886,12 +885,12 @@ void inet_csk_listen_stop(struct sock *sk)
 		sk_acceptq_removed(sk);
 		reqsk_put(req);
 	}
-	if (queue->fastopenq) {
+	if (queue->fastopenq.rskq_rst_head) {
 		/* Free all the reqs queued in rskq_rst_head. */
-		spin_lock_bh(&queue->fastopenq->lock);
-		acc_req = queue->fastopenq->rskq_rst_head;
-		queue->fastopenq->rskq_rst_head = NULL;
-		spin_unlock_bh(&queue->fastopenq->lock);
+		spin_lock_bh(&queue->fastopenq.lock);
+		acc_req = queue->fastopenq.rskq_rst_head;
+		queue->fastopenq.rskq_rst_head = NULL;
+		spin_unlock_bh(&queue->fastopenq.lock);
 		while ((req = acc_req) != NULL) {
 			acc_req = req->dl_next;
 			reqsk_put(req);
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index b8b8fa184f75..3c96fa87ff9e 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2253,13 +2253,6 @@ int tcp_disconnect(struct sock *sk, int flags)
 }
 EXPORT_SYMBOL(tcp_disconnect);
 
-void tcp_sock_destruct(struct sock *sk)
-{
-	inet_sock_destruct(sk);
-
-	kfree(inet_csk(sk)->icsk_accept_queue.fastopenq);
-}
-
 static inline bool tcp_can_repair_sock(const struct sock *sk)
 {
 	return ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN) &&
@@ -2581,7 +2574,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
 		    TCPF_LISTEN))) {
 			tcp_fastopen_init_key_once(true);
 
-			err = fastopen_init_queue(sk, val);
+			fastopen_queue_tune(sk, val);
 		} else {
 			err = -EINVAL;
 		}
@@ -2849,10 +2842,7 @@ static int do_tcp_getsockopt(struct sock *sk, int level,
 		break;
 
 	case TCP_FASTOPEN:
-		if (icsk->icsk_accept_queue.fastopenq)
-			val = icsk->icsk_accept_queue.fastopenq->max_qlen;
-		else
-			val = 0;
+		val = icsk->icsk_accept_queue.fastopenq.max_qlen;
 		break;
 
 	case TCP_TIMESTAMP:
diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
index db43c6286cf7..f69f436fcbcc 100644
--- a/net/ipv4/tcp_fastopen.c
+++ b/net/ipv4/tcp_fastopen.c
@@ -142,9 +142,9 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 	if (!child)
 		return NULL;
 
-	spin_lock(&queue->fastopenq->lock);
-	queue->fastopenq->qlen++;
-	spin_unlock(&queue->fastopenq->lock);
+	spin_lock(&queue->fastopenq.lock);
+	queue->fastopenq.qlen++;
+	spin_unlock(&queue->fastopenq.lock);
 
 	/* Initialize the child socket. Have to fix some values to take
 	 * into account the child is a Fast Open socket and is created
@@ -237,8 +237,8 @@ static bool tcp_fastopen_queue_check(struct sock *sk)
 	 * between qlen overflow causing Fast Open to be disabled
 	 * temporarily vs a server not supporting Fast Open at all.
 	 */
-	fastopenq = inet_csk(sk)->icsk_accept_queue.fastopenq;
-	if (!fastopenq || fastopenq->max_qlen == 0)
+	fastopenq = &inet_csk(sk)->icsk_accept_queue.fastopenq;
+	if (fastopenq->max_qlen == 0)
 		return false;
 
 	if (fastopenq->qlen >= fastopenq->max_qlen) {
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index f551e9e862db..64ece718d66c 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2186,7 +2186,7 @@ static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i)
 	const struct tcp_sock *tp = tcp_sk(sk);
 	const struct inet_connection_sock *icsk = inet_csk(sk);
 	const struct inet_sock *inet = inet_sk(sk);
-	struct fastopen_queue *fastopenq = icsk->icsk_accept_queue.fastopenq;
+	const struct fastopen_queue *fastopenq = &icsk->icsk_accept_queue.fastopenq;
 	__be32 dest = inet->inet_daddr;
 	__be32 src = inet->inet_rcv_saddr;
 	__u16 destp = ntohs(inet->inet_dport);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 97bc26e0cd0f..0ac64f47f882 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1672,7 +1672,7 @@ static void get_tcp6_sock(struct seq_file *seq, struct sock *sp, int i)
 	const struct inet_sock *inet = inet_sk(sp);
 	const struct tcp_sock *tp = tcp_sk(sp);
 	const struct inet_connection_sock *icsk = inet_csk(sp);
-	struct fastopen_queue *fastopenq = icsk->icsk_accept_queue.fastopenq;
+	const struct fastopen_queue *fastopenq = &icsk->icsk_accept_queue.fastopenq;
 
 	dest  = &sp->sk_v6_daddr;
 	src   = &sp->sk_v6_rcv_saddr;
@@ -1716,7 +1716,7 @@ static void get_tcp6_sock(struct seq_file *seq, struct sock *sp, int i)
 		   (icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
 		   tp->snd_cwnd,
 		   sp->sk_state == TCP_LISTEN ?
-			(fastopenq ? fastopenq->max_qlen : 0) :
+			fastopenq->max_qlen :
 			(tcp_in_initial_slowstart(tp) ? -1 : tp->snd_ssthresh)
 		   );
 }
-- 
cgit v1.2.3


From ca6fb06518836ef9b65dc0aac02ff97704d52a05 Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Fri, 2 Oct 2015 11:43:35 -0700
Subject: tcp: attach SYNACK messages to request sockets instead of listener

If a listen backlog is very big (to avoid syncookies), then
the listener sk->sk_wmem_alloc is the main source of false
sharing, as we need to touch it twice per SYNACK re-transmit
and TX completion.

(One SYN packet takes listener lock once, but up to 6 SYNACK
are generated)

By attaching the skb to the request socket, we remove this
source of contention.

Tested:

 listen(fd, 10485760); // single listener (no SO_REUSEPORT)
 16 RX/TX queue NIC
 Sustain a SYNFLOOD attack of ~320,000 SYN per second,
 Sending ~1,400,000 SYNACK per second.
 Perf profiles now show listener spinlock being next bottleneck.

    20.29%  [kernel]  [k] queued_spin_lock_slowpath
    10.06%  [kernel]  [k] __inet_lookup_established
     5.12%  [kernel]  [k] reqsk_timer_handler
     3.22%  [kernel]  [k] get_next_timer_interrupt
     3.00%  [kernel]  [k] tcp_make_synack
     2.77%  [kernel]  [k] ipt_do_table
     2.70%  [kernel]  [k] run_timer_softirq
     2.50%  [kernel]  [k] ip_finish_output
     2.04%  [kernel]  [k] cascade

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/net/tcp.h               |  6 ++++--
 net/ipv4/inet_connection_sock.c |  2 +-
 net/ipv4/tcp_fastopen.c         |  4 ++--
 net/ipv4/tcp_input.c            | 23 ++++++++++++-----------
 net/ipv4/tcp_ipv4.c             |  5 +++--
 net/ipv4/tcp_output.c           | 22 +++++++++++++++-------
 net/ipv6/tcp_ipv6.c             |  5 +++--
 net/sched/sch_fq.c              | 12 +++++++-----
 8 files changed, 47 insertions(+), 32 deletions(-)

(limited to 'net/ipv4/tcp_fastopen.c')

diff --git a/include/net/tcp.h b/include/net/tcp.h
index 225e9561af35..a6be56d5f0e3 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -462,7 +462,8 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
 int tcp_connect(struct sock *sk);
 struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
 				struct request_sock *req,
-				struct tcp_fastopen_cookie *foc);
+				struct tcp_fastopen_cookie *foc,
+				bool attach_req);
 int tcp_disconnect(struct sock *sk, int flags);
 
 void tcp_finish_connect(struct sock *sk, struct sk_buff *skb);
@@ -1715,7 +1716,8 @@ struct tcp_request_sock_ops {
 	__u32 (*init_seq)(const struct sk_buff *skb);
 	int (*send_synack)(const struct sock *sk, struct dst_entry *dst,
 			   struct flowi *fl, struct request_sock *req,
-			   u16 queue_mapping, struct tcp_fastopen_cookie *foc);
+			   u16 queue_mapping, struct tcp_fastopen_cookie *foc,
+			   bool attach_req);
 };
 
 #ifdef CONFIG_SYN_COOKIES
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 80904df02187..099e0ea9242a 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -628,7 +628,7 @@ static void reqsk_queue_hash_req(struct request_sock *req,
 	 * are committed to memory and refcnt initialized.
 	 */
 	smp_wmb();
-	atomic_set(&req->rsk_refcnt, 2);
+	atomic_set(&req->rsk_refcnt, 2 + 1);
 }
 
 void inet_csk_reqsk_queue_hash_add(struct sock *sk, struct request_sock *req,
diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
index f69f436fcbcc..410ac481fda0 100644
--- a/net/ipv4/tcp_fastopen.c
+++ b/net/ipv4/tcp_fastopen.c
@@ -161,13 +161,13 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 	tp->snd_wnd = ntohs(tcp_hdr(skb)->window);
 
 	/* Activate the retrans timer so that SYNACK can be retransmitted.
-	 * The request socket is not added to the SYN table of the parent
+	 * The request socket is not added to the ehash
 	 * because it's been added to the accept queue directly.
 	 */
 	inet_csk_reset_xmit_timer(child, ICSK_TIME_RETRANS,
 				  TCP_TIMEOUT_INIT, TCP_RTO_MAX);
 
-	atomic_set(&req->rsk_refcnt, 1);
+	atomic_set(&req->rsk_refcnt, 2);
 	/* Add the child socket directly into the accept queue */
 	inet_csk_reqsk_queue_add(sk, req, child);
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index a56912772354..27108757c310 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -6120,8 +6120,6 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
 	struct request_sock *req;
 	bool want_cookie = false;
 	struct flowi fl;
-	int err;
-
 
 	/* TW buckets are converted to open requests without
 	 * limitations, they conserve resources and peer is
@@ -6230,21 +6228,24 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
 	tcp_rsk(req)->snt_isn = isn;
 	tcp_rsk(req)->txhash = net_tx_rndhash();
 	tcp_openreq_init_rwin(req, sk, dst);
-	if (!want_cookie)
+	if (!want_cookie) {
 		fastopen_sk = tcp_try_fastopen(sk, skb, req, &foc, dst);
-	err = af_ops->send_synack(fastopen_sk ?: sk, dst, &fl, req,
-				  skb_get_queue_mapping(skb), &foc);
+		tcp_reqsk_record_syn(sk, req, skb);
+	}
 	if (fastopen_sk) {
+		af_ops->send_synack(fastopen_sk, dst, &fl, req,
+				    skb_get_queue_mapping(skb), &foc, false);
 		sock_put(fastopen_sk);
 	} else {
-		if (err || want_cookie)
-			goto drop_and_free;
-
 		tcp_rsk(req)->tfo_listener = false;
-		inet_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
+		if (!want_cookie)
+			inet_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
+		af_ops->send_synack(sk, dst, &fl, req,
+				    skb_get_queue_mapping(skb), &foc, !want_cookie);
+		if (want_cookie)
+			goto drop_and_free;
 	}
-	tcp_reqsk_record_syn(sk, req, skb);
-
+	reqsk_put(req);
 	return 0;
 
 drop_and_release:
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index bfe9d39ee87d..ac2ea73e9aaf 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -822,7 +822,8 @@ static int tcp_v4_send_synack(const struct sock *sk, struct dst_entry *dst,
 			      struct flowi *fl,
 			      struct request_sock *req,
 			      u16 queue_mapping,
-			      struct tcp_fastopen_cookie *foc)
+			      struct tcp_fastopen_cookie *foc,
+				  bool attach_req)
 {
 	const struct inet_request_sock *ireq = inet_rsk(req);
 	struct flowi4 fl4;
@@ -833,7 +834,7 @@ static int tcp_v4_send_synack(const struct sock *sk, struct dst_entry *dst,
 	if (!dst && (dst = inet_csk_route_req(sk, &fl4, req)) == NULL)
 		return -1;
 
-	skb = tcp_make_synack(sk, dst, req, foc);
+	skb = tcp_make_synack(sk, dst, req, foc, attach_req);
 
 	if (skb) {
 		__tcp_v4_send_check(skb, ireq->ir_loc_addr, ireq->ir_rmt_addr);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 09bb082ca1a7..55ed3266b05f 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2947,7 +2947,8 @@ int tcp_send_synack(struct sock *sk)
  */
 struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
 				struct request_sock *req,
-				struct tcp_fastopen_cookie *foc)
+				struct tcp_fastopen_cookie *foc,
+				bool attach_req)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
 	const struct tcp_sock *tp = tcp_sk(sk);
@@ -2959,11 +2960,7 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
 	u16 user_mss;
 	int mss;
 
-	/* sk is a const pointer, because we want to express multiple cpus
-	 * might call us concurrently.
-	 * sock_wmalloc() will change sk->sk_wmem_alloc in an atomic way.
-	 */
-	skb = sock_wmalloc((struct sock *)sk, MAX_TCP_HEADER, 1, GFP_ATOMIC);
+	skb = alloc_skb(MAX_TCP_HEADER, GFP_ATOMIC);
 	if (unlikely(!skb)) {
 		dst_release(dst);
 		return NULL;
@@ -2971,6 +2968,17 @@ struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
 	/* Reserve space for headers. */
 	skb_reserve(skb, MAX_TCP_HEADER);
 
+	if (attach_req) {
+		skb->destructor = sock_edemux;
+		sock_hold(req_to_sk(req));
+		skb->sk = req_to_sk(req);
+	} else {
+		/* sk is a const pointer, because we want to express multiple
+		 * cpu might call us concurrently.
+		 * sk->sk_wmem_alloc in an atomic, we can promote to rw.
+		 */
+		skb_set_owner_w(skb, (struct sock *)sk);
+	}
 	skb_dst_set(skb, dst);
 
 	mss = dst_metric_advmss(dst);
@@ -3510,7 +3518,7 @@ int tcp_rtx_synack(const struct sock *sk, struct request_sock *req)
 	int res;
 
 	tcp_rsk(req)->txhash = net_tx_rndhash();
-	res = af_ops->send_synack(sk, NULL, &fl, req, 0, NULL);
+	res = af_ops->send_synack(sk, NULL, &fl, req, 0, NULL, true);
 	if (!res) {
 		TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_RETRANSSEGS);
 		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSYNRETRANS);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index a215614cfb2b..3d18571811c5 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -438,7 +438,8 @@ static int tcp_v6_send_synack(const struct sock *sk, struct dst_entry *dst,
 			      struct flowi *fl,
 			      struct request_sock *req,
 			      u16 queue_mapping,
-			      struct tcp_fastopen_cookie *foc)
+			      struct tcp_fastopen_cookie *foc,
+			      bool attach_req)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
 	struct ipv6_pinfo *np = inet6_sk(sk);
@@ -451,7 +452,7 @@ static int tcp_v6_send_synack(const struct sock *sk, struct dst_entry *dst,
 					       IPPROTO_TCP)) == NULL)
 		goto done;
 
-	skb = tcp_make_synack(sk, dst, req, foc);
+	skb = tcp_make_synack(sk, dst, req, foc, attach_req);
 
 	if (skb) {
 		__tcp_v6_send_check(skb, &ireq->ir_v6_loc_addr,
diff --git a/net/sched/sch_fq.c b/net/sched/sch_fq.c
index f377702d4b91..3386cce4751e 100644
--- a/net/sched/sch_fq.c
+++ b/net/sched/sch_fq.c
@@ -224,13 +224,15 @@ static struct fq_flow *fq_classify(struct sk_buff *skb, struct fq_sched_data *q)
 	if (unlikely((skb->priority & TC_PRIO_MAX) == TC_PRIO_CONTROL))
 		return &q->internal;
 
-	/* SYNACK messages are attached to a listener socket.
-	 * 1) They are not part of a 'flow' yet
-	 * 2) We do not want to rate limit them (eg SYNFLOOD attack),
+	/* SYNACK messages are attached to a TCP_NEW_SYN_RECV request socket
+	 * 1) request sockets are not full blown,
+	 *    they do not contain sk_pacing_rate
+	 * 2) They are not part of a 'flow' yet
+	 * 3) We do not want to rate limit them (eg SYNFLOOD attack),
 	 *    especially if the listener set SO_MAX_PACING_RATE
-	 * 3) We pretend they are orphaned
+	 * 4) We pretend they are orphaned
 	 */
-	if (!sk || sk->sk_state == TCP_LISTEN) {
+	if (!sk || sk->sk_state == TCP_NEW_SYN_RECV) {
 		unsigned long hash = skb_get_hash(skb) & q->orphan_mask;
 
 		/* By forcing low order bit to 1, we make sure to not
-- 
cgit v1.2.3


From 7656d842de93fd2d2de7b403062cad757cadf1df Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Sun, 4 Oct 2015 21:08:07 -0700
Subject: tcp: fix fastopen races vs lockless listener

There are multiple races that need fixes :

1) skb_get() + queue skb + kfree_skb() is racy

An accept() can be done on another cpu, data consumed immediately.
tcp_recvmsg() uses __kfree_skb() as it is assumed all skb found in
socket receive queue are private.

Then the kfree_skb() in tcp_rcv_state_process() uses an already freed skb

2) tcp_reqsk_record_syn() needs to be done before tcp_try_fastopen()
for the same reasons.

3) We want to send the SYNACK before queueing child into accept queue,
otherwise we might reintroduce the ooo issue fixed in
commit 7c85af881044 ("tcp: avoid reorders for TFO passive connections")

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 net/ipv4/tcp_fastopen.c | 26 +++++++-------------------
 net/ipv4/tcp_input.c    |  6 +++++-
 2 files changed, 12 insertions(+), 20 deletions(-)

(limited to 'net/ipv4/tcp_fastopen.c')

diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
index 410ac481fda0..93396bf7b475 100644
--- a/net/ipv4/tcp_fastopen.c
+++ b/net/ipv4/tcp_fastopen.c
@@ -168,8 +168,6 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 				  TCP_TIMEOUT_INIT, TCP_RTO_MAX);
 
 	atomic_set(&req->rsk_refcnt, 2);
-	/* Add the child socket directly into the accept queue */
-	inet_csk_reqsk_queue_add(sk, req, child);
 
 	/* Now finish processing the fastopen child socket. */
 	inet_csk(child)->icsk_af_ops->rebuild_header(child);
@@ -178,12 +176,10 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 	tcp_init_metrics(child);
 	tcp_init_buffer_space(child);
 
-	/* Queue the data carried in the SYN packet. We need to first
-	 * bump skb's refcnt because the caller will attempt to free it.
-	 * Note that IPv6 might also have used skb_get() trick
-	 * in tcp_v6_conn_request() to keep this SYN around (treq->pktopts)
-	 * So we need to eventually get a clone of the packet,
-	 * before inserting it in sk_receive_queue.
+	/* Queue the data carried in the SYN packet.
+	 * We used to play tricky games with skb_get().
+	 * With lockless listener, it is a dead end.
+	 * Do not think about it.
 	 *
 	 * XXX (TFO) - we honor a zero-payload TFO request for now,
 	 * (any reason not to?) but no need to queue the skb since
@@ -191,12 +187,7 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 	 */
 	end_seq = TCP_SKB_CB(skb)->end_seq;
 	if (end_seq != TCP_SKB_CB(skb)->seq + 1) {
-		struct sk_buff *skb2;
-
-		if (unlikely(skb_shared(skb)))
-			skb2 = skb_clone(skb, GFP_ATOMIC);
-		else
-			skb2 = skb_get(skb);
+		struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
 
 		if (likely(skb2)) {
 			skb_dst_drop(skb2);
@@ -214,12 +205,9 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 		}
 	}
 	tcp_rsk(req)->rcv_nxt = tp->rcv_nxt = end_seq;
-	sk->sk_data_ready(sk);
-	bh_unlock_sock(child);
-	/* Note: sock_put(child) will be done by tcp_conn_request()
-	 * after SYNACK packet is sent.
+	/* tcp_conn_request() is sending the SYNACK,
+	 * and queues the child into listener accept queue.
 	 */
-	WARN_ON(!req->sk);
 	return child;
 }
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 27108757c310..a95c8eb04ff7 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -6229,12 +6229,16 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
 	tcp_rsk(req)->txhash = net_tx_rndhash();
 	tcp_openreq_init_rwin(req, sk, dst);
 	if (!want_cookie) {
-		fastopen_sk = tcp_try_fastopen(sk, skb, req, &foc, dst);
 		tcp_reqsk_record_syn(sk, req, skb);
+		fastopen_sk = tcp_try_fastopen(sk, skb, req, &foc, dst);
 	}
 	if (fastopen_sk) {
 		af_ops->send_synack(fastopen_sk, dst, &fl, req,
 				    skb_get_queue_mapping(skb), &foc, false);
+		/* Add the child socket directly into the accept queue */
+		inet_csk_reqsk_queue_add(sk, req, fastopen_sk);
+		sk->sk_data_ready(sk);
+		bh_unlock_sock(fastopen_sk);
 		sock_put(fastopen_sk);
 	} else {
 		tcp_rsk(req)->tfo_listener = false;
-- 
cgit v1.2.3


From 5e0724d027f0548511a2165a209572d48fe7a4c8 Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Thu, 22 Oct 2015 08:20:46 -0700
Subject: tcp/dccp: fix hashdance race for passive sessions

Multiple cpus can process duplicates of incoming ACK messages
matching a SYN_RECV request socket. This is a rare event under
normal operations, but definitely can happen.

Only one must win the race, otherwise corruption would occur.

To fix this without adding new atomic ops, we use logic in
inet_ehash_nolisten() to detect the request was present in the same
ehash bucket where we try to insert the new child.

If request socket was not found, we have to undo the child creation.

This actually removes a spin_lock()/spin_unlock() pair in
reqsk_queue_unlink() for the fast path.

Fixes: e994b2f0fb92 ("tcp: do not lock listener to process SYN packets")
Fixes: 079096f103fa ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/net/inet_connection_sock.h |  7 ++++++-
 include/net/inet_hashtables.h      |  4 ++--
 include/net/tcp.h                  |  4 +++-
 net/dccp/dccp.h                    |  4 +++-
 net/dccp/ipv4.c                    |  6 ++++--
 net/dccp/ipv6.c                    |  9 ++++++---
 net/dccp/minisocks.c               | 14 +++++++-------
 net/ipv4/inet_connection_sock.c    | 33 +++++++++++++++++++++++++-------
 net/ipv4/inet_hashtables.c         | 39 ++++++++++++++++++++++++--------------
 net/ipv4/syncookies.c              |  4 +++-
 net/ipv4/tcp_fastopen.c            |  4 +++-
 net/ipv4/tcp_ipv4.c                |  6 ++++--
 net/ipv4/tcp_minisocks.c           | 11 ++++-------
 net/ipv6/tcp_ipv6.c                |  9 ++++++---
 14 files changed, 102 insertions(+), 52 deletions(-)

(limited to 'net/ipv4/tcp_fastopen.c')

diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index 63615709839d..481fe1c9044c 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -43,7 +43,9 @@ struct inet_connection_sock_af_ops {
 	int	    (*conn_request)(struct sock *sk, struct sk_buff *skb);
 	struct sock *(*syn_recv_sock)(const struct sock *sk, struct sk_buff *skb,
 				      struct request_sock *req,
-				      struct dst_entry *dst);
+				      struct dst_entry *dst,
+				      struct request_sock *req_unhash,
+				      bool *own_req);
 	u16	    net_header_len;
 	u16	    net_frag_header_len;
 	u16	    sockaddr_len;
@@ -272,6 +274,9 @@ void inet_csk_reqsk_queue_add(struct sock *sk, struct request_sock *req,
 			      struct sock *child);
 void inet_csk_reqsk_queue_hash_add(struct sock *sk, struct request_sock *req,
 				   unsigned long timeout);
+struct sock *inet_csk_complete_hashdance(struct sock *sk, struct sock *child,
+					 struct request_sock *req,
+					 bool own_req);
 
 static inline void inet_csk_reqsk_queue_added(struct sock *sk)
 {
diff --git a/include/net/inet_hashtables.h b/include/net/inet_hashtables.h
index 6683ada25fef..de2e3ade6102 100644
--- a/include/net/inet_hashtables.h
+++ b/include/net/inet_hashtables.h
@@ -205,8 +205,8 @@ void inet_put_port(struct sock *sk);
 
 void inet_hashinfo_init(struct inet_hashinfo *h);
 
-int inet_ehash_insert(struct sock *sk, struct sock *osk);
-void __inet_hash_nolisten(struct sock *sk, struct sock *osk);
+bool inet_ehash_insert(struct sock *sk, struct sock *osk);
+bool inet_ehash_nolisten(struct sock *sk, struct sock *osk);
 void __inet_hash(struct sock *sk, struct sock *osk);
 void inet_hash(struct sock *sk);
 void inet_unhash(struct sock *sk);
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 11e320412216..f80e74c5ad18 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -457,7 +457,9 @@ struct sock *tcp_create_openreq_child(const struct sock *sk,
 void tcp_ca_openreq_child(struct sock *sk, const struct dst_entry *dst);
 struct sock *tcp_v4_syn_recv_sock(const struct sock *sk, struct sk_buff *skb,
 				  struct request_sock *req,
-				  struct dst_entry *dst);
+				  struct dst_entry *dst,
+				  struct request_sock *req_unhash,
+				  bool *own_req);
 int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb);
 int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
 int tcp_connect(struct sock *sk);
diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h
index 923f5a180134..b0e28d24e1a7 100644
--- a/net/dccp/dccp.h
+++ b/net/dccp/dccp.h
@@ -278,7 +278,9 @@ int dccp_v4_do_rcv(struct sock *sk, struct sk_buff *skb);
 
 struct sock *dccp_v4_request_recv_sock(const struct sock *sk, struct sk_buff *skb,
 				       struct request_sock *req,
-				       struct dst_entry *dst);
+				       struct dst_entry *dst,
+				       struct request_sock *req_unhash,
+				       bool *own_req);
 struct sock *dccp_check_req(struct sock *sk, struct sk_buff *skb,
 			    struct request_sock *req);
 
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 59bc180b02d8..5684e14932bd 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -393,7 +393,9 @@ static inline u64 dccp_v4_init_sequence(const struct sk_buff *skb)
 struct sock *dccp_v4_request_recv_sock(const struct sock *sk,
 				       struct sk_buff *skb,
 				       struct request_sock *req,
-				       struct dst_entry *dst)
+				       struct dst_entry *dst,
+				       struct request_sock *req_unhash,
+				       bool *own_req)
 {
 	struct inet_request_sock *ireq;
 	struct inet_sock *newinet;
@@ -426,7 +428,7 @@ struct sock *dccp_v4_request_recv_sock(const struct sock *sk,
 
 	if (__inet_inherit_port(sk, newsk) < 0)
 		goto put_and_exit;
-	__inet_hash_nolisten(newsk, NULL);
+	*own_req = inet_ehash_nolisten(newsk, req_to_sk(req_unhash));
 
 	return newsk;
 
diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c
index d9cc731f2619..ef4e48ce9143 100644
--- a/net/dccp/ipv6.c
+++ b/net/dccp/ipv6.c
@@ -380,7 +380,9 @@ drop:
 static struct sock *dccp_v6_request_recv_sock(const struct sock *sk,
 					      struct sk_buff *skb,
 					      struct request_sock *req,
-					      struct dst_entry *dst)
+					      struct dst_entry *dst,
+					      struct request_sock *req_unhash,
+					      bool *own_req)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
 	struct ipv6_pinfo *newnp;
@@ -393,7 +395,8 @@ static struct sock *dccp_v6_request_recv_sock(const struct sock *sk,
 		/*
 		 *	v6 mapped
 		 */
-		newsk = dccp_v4_request_recv_sock(sk, skb, req, dst);
+		newsk = dccp_v4_request_recv_sock(sk, skb, req, dst,
+						  req_unhash, own_req);
 		if (newsk == NULL)
 			return NULL;
 
@@ -511,7 +514,7 @@ static struct sock *dccp_v6_request_recv_sock(const struct sock *sk,
 		dccp_done(newsk);
 		goto out;
 	}
-	__inet_hash(newsk, NULL);
+	*own_req = inet_ehash_nolisten(newsk, req_to_sk(req_unhash));
 
 	return newsk;
 
diff --git a/net/dccp/minisocks.c b/net/dccp/minisocks.c
index d10aace43672..1994f8af646b 100644
--- a/net/dccp/minisocks.c
+++ b/net/dccp/minisocks.c
@@ -143,6 +143,7 @@ struct sock *dccp_check_req(struct sock *sk, struct sk_buff *skb,
 {
 	struct sock *child = NULL;
 	struct dccp_request_sock *dreq = dccp_rsk(req);
+	bool own_req;
 
 	/* Check for retransmitted REQUEST */
 	if (dccp_hdr(skb)->dccph_type == DCCP_PKT_REQUEST) {
@@ -182,14 +183,13 @@ struct sock *dccp_check_req(struct sock *sk, struct sk_buff *skb,
 	if (dccp_parse_options(sk, dreq, skb))
 		 goto drop;
 
-	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
-	if (child == NULL)
+	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL,
+							 req, &own_req);
+	if (!child)
 		goto listen_overflow;
 
-	inet_csk_reqsk_queue_drop(sk, req);
-	inet_csk_reqsk_queue_add(sk, req, child);
-out:
-	return child;
+	return inet_csk_complete_hashdance(sk, child, req, own_req);
+
 listen_overflow:
 	dccp_pr_debug("listen_overflow!\n");
 	DCCP_SKB_CB(skb)->dccpd_reset_code = DCCP_RESET_CODE_TOO_BUSY;
@@ -198,7 +198,7 @@ drop:
 		req->rsk_ops->send_reset(sk, skb);
 
 	inet_csk_reqsk_queue_drop(sk, req);
-	goto out;
+	return NULL;
 }
 
 EXPORT_SYMBOL_GPL(dccp_check_req);
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 8430bc8ccd58..1feb15f23de8 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -523,15 +523,15 @@ static bool reqsk_queue_unlink(struct request_sock_queue *queue,
 			       struct request_sock *req)
 {
 	struct inet_hashinfo *hashinfo = req_to_sk(req)->sk_prot->h.hashinfo;
-	spinlock_t *lock;
-	bool found;
+	bool found = false;
 
-	lock = inet_ehash_lockp(hashinfo, req->rsk_hash);
-
-	spin_lock(lock);
-	found = __sk_nulls_del_node_init_rcu(req_to_sk(req));
-	spin_unlock(lock);
+	if (sk_hashed(req_to_sk(req))) {
+		spinlock_t *lock = inet_ehash_lockp(hashinfo, req->rsk_hash);
 
+		spin_lock(lock);
+		found = __sk_nulls_del_node_init_rcu(req_to_sk(req));
+		spin_unlock(lock);
+	}
 	if (timer_pending(&req->rsk_timer) && del_timer_sync(&req->rsk_timer))
 		reqsk_put(req);
 	return found;
@@ -811,6 +811,25 @@ void inet_csk_reqsk_queue_add(struct sock *sk, struct request_sock *req,
 }
 EXPORT_SYMBOL(inet_csk_reqsk_queue_add);
 
+struct sock *inet_csk_complete_hashdance(struct sock *sk, struct sock *child,
+					 struct request_sock *req, bool own_req)
+{
+	if (own_req) {
+		inet_csk_reqsk_queue_drop(sk, req);
+		reqsk_queue_removed(&inet_csk(sk)->icsk_accept_queue, req);
+		inet_csk_reqsk_queue_add(sk, req, child);
+		/* Warning: caller must not call reqsk_put(req);
+		 * child stole last reference on it.
+		 */
+		return child;
+	}
+	/* Too bad, another child took ownership of the request, undo. */
+	bh_unlock_sock(child);
+	sock_put(child);
+	return NULL;
+}
+EXPORT_SYMBOL(inet_csk_complete_hashdance);
+
 /*
  *	This routine closes sockets which have been at least partially
  *	opened, but not yet accepted.
diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c
index 958728a22001..ccc5980797fc 100644
--- a/net/ipv4/inet_hashtables.c
+++ b/net/ipv4/inet_hashtables.c
@@ -407,13 +407,13 @@ static u32 inet_sk_port_offset(const struct sock *sk)
 /* insert a socket into ehash, and eventually remove another one
  * (The another one can be a SYN_RECV or TIMEWAIT
  */
-int inet_ehash_insert(struct sock *sk, struct sock *osk)
+bool inet_ehash_insert(struct sock *sk, struct sock *osk)
 {
 	struct inet_hashinfo *hashinfo = sk->sk_prot->h.hashinfo;
 	struct hlist_nulls_head *list;
 	struct inet_ehash_bucket *head;
 	spinlock_t *lock;
-	int ret = 0;
+	bool ret = true;
 
 	WARN_ON_ONCE(!sk_unhashed(sk));
 
@@ -423,30 +423,41 @@ int inet_ehash_insert(struct sock *sk, struct sock *osk)
 	lock = inet_ehash_lockp(hashinfo, sk->sk_hash);
 
 	spin_lock(lock);
-	__sk_nulls_add_node_rcu(sk, list);
 	if (osk) {
-		WARN_ON(sk->sk_hash != osk->sk_hash);
-		sk_nulls_del_node_init_rcu(osk);
+		WARN_ON_ONCE(sk->sk_hash != osk->sk_hash);
+		ret = sk_nulls_del_node_init_rcu(osk);
 	}
+	if (ret)
+		__sk_nulls_add_node_rcu(sk, list);
 	spin_unlock(lock);
 	return ret;
 }
 
-void __inet_hash_nolisten(struct sock *sk, struct sock *osk)
+bool inet_ehash_nolisten(struct sock *sk, struct sock *osk)
 {
-	inet_ehash_insert(sk, osk);
-	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
+	bool ok = inet_ehash_insert(sk, osk);
+
+	if (ok) {
+		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
+	} else {
+		percpu_counter_inc(sk->sk_prot->orphan_count);
+		sk->sk_state = TCP_CLOSE;
+		sock_set_flag(sk, SOCK_DEAD);
+		inet_csk_destroy_sock(sk);
+	}
+	return ok;
 }
-EXPORT_SYMBOL_GPL(__inet_hash_nolisten);
+EXPORT_SYMBOL_GPL(inet_ehash_nolisten);
 
 void __inet_hash(struct sock *sk, struct sock *osk)
 {
 	struct inet_hashinfo *hashinfo = sk->sk_prot->h.hashinfo;
 	struct inet_listen_hashbucket *ilb;
 
-	if (sk->sk_state != TCP_LISTEN)
-		return __inet_hash_nolisten(sk, osk);
-
+	if (sk->sk_state != TCP_LISTEN) {
+		inet_ehash_nolisten(sk, osk);
+		return;
+	}
 	WARN_ON(!sk_unhashed(sk));
 	ilb = &hashinfo->listening_hash[inet_sk_listen_hashfn(sk)];
 
@@ -567,7 +578,7 @@ ok:
 		inet_bind_hash(sk, tb, port);
 		if (sk_unhashed(sk)) {
 			inet_sk(sk)->inet_sport = htons(port);
-			__inet_hash_nolisten(sk, (struct sock *)tw);
+			inet_ehash_nolisten(sk, (struct sock *)tw);
 		}
 		if (tw)
 			inet_twsk_bind_unhash(tw, hinfo);
@@ -584,7 +595,7 @@ ok:
 	tb  = inet_csk(sk)->icsk_bind_hash;
 	spin_lock_bh(&head->lock);
 	if (sk_head(&tb->owners) == sk && !sk->sk_bind_node.next) {
-		__inet_hash_nolisten(sk, NULL);
+		inet_ehash_nolisten(sk, NULL);
 		spin_unlock_bh(&head->lock);
 		return 0;
 	} else {
diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index 4c0892badb8b..4cbe9f0a4281 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -221,8 +221,10 @@ struct sock *tcp_get_cookie_sock(struct sock *sk, struct sk_buff *skb,
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct sock *child;
+	bool own_req;
 
-	child = icsk->icsk_af_ops->syn_recv_sock(sk, skb, req, dst);
+	child = icsk->icsk_af_ops->syn_recv_sock(sk, skb, req, dst,
+						 NULL, &own_req);
 	if (child) {
 		atomic_set(&req->rsk_refcnt, 1);
 		sock_rps_save_rxhash(child, skb);
diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c
index 93396bf7b475..55be6ac70cff 100644
--- a/net/ipv4/tcp_fastopen.c
+++ b/net/ipv4/tcp_fastopen.c
@@ -133,12 +133,14 @@ static struct sock *tcp_fastopen_create_child(struct sock *sk,
 	struct request_sock_queue *queue = &inet_csk(sk)->icsk_accept_queue;
 	struct sock *child;
 	u32 end_seq;
+	bool own_req;
 
 	req->num_retrans = 0;
 	req->num_timeout = 0;
 	req->sk = NULL;
 
-	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
+	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL,
+							 NULL, &own_req);
 	if (!child)
 		return NULL;
 
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 30dd45c1f568..1c2648bbac4b 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1247,7 +1247,9 @@ EXPORT_SYMBOL(tcp_v4_conn_request);
  */
 struct sock *tcp_v4_syn_recv_sock(const struct sock *sk, struct sk_buff *skb,
 				  struct request_sock *req,
-				  struct dst_entry *dst)
+				  struct dst_entry *dst,
+				  struct request_sock *req_unhash,
+				  bool *own_req)
 {
 	struct inet_request_sock *ireq;
 	struct inet_sock *newinet;
@@ -1323,7 +1325,7 @@ struct sock *tcp_v4_syn_recv_sock(const struct sock *sk, struct sk_buff *skb,
 
 	if (__inet_inherit_port(sk, newsk) < 0)
 		goto put_and_exit;
-	__inet_hash_nolisten(newsk, NULL);
+	*own_req = inet_ehash_nolisten(newsk, req_to_sk(req_unhash));
 
 	return newsk;
 
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 1fd5d413a664..3575dd1e5b67 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -580,6 +580,7 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
 	const struct tcphdr *th = tcp_hdr(skb);
 	__be32 flg = tcp_flag_word(th) & (TCP_FLAG_RST|TCP_FLAG_SYN|TCP_FLAG_ACK);
 	bool paws_reject = false;
+	bool own_req;
 
 	tmp_opt.saw_tstamp = 0;
 	if (th->doff > (sizeof(struct tcphdr)>>2)) {
@@ -767,18 +768,14 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
 	 * ESTABLISHED STATE. If it will be dropped after
 	 * socket is created, wait for troubles.
 	 */
-	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
+	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL,
+							 req, &own_req);
 	if (!child)
 		goto listen_overflow;
 
 	sock_rps_save_rxhash(child, skb);
 	tcp_synack_rtt_meas(child, req);
-	inet_csk_reqsk_queue_drop(sk, req);
-	inet_csk_reqsk_queue_add(sk, req, child);
-	/* Warning: caller must not call reqsk_put(req);
-	 * child stole last reference on it.
-	 */
-	return child;
+	return inet_csk_complete_hashdance(sk, child, req, own_req);
 
 listen_overflow:
 	if (!sysctl_tcp_abort_on_overflow) {
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index f495d189f5e0..714bc5ad096e 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -965,7 +965,9 @@ drop:
 
 static struct sock *tcp_v6_syn_recv_sock(const struct sock *sk, struct sk_buff *skb,
 					 struct request_sock *req,
-					 struct dst_entry *dst)
+					 struct dst_entry *dst,
+					 struct request_sock *req_unhash,
+					 bool *own_req)
 {
 	struct inet_request_sock *ireq;
 	struct ipv6_pinfo *newnp;
@@ -984,7 +986,8 @@ static struct sock *tcp_v6_syn_recv_sock(const struct sock *sk, struct sk_buff *
 		 *	v6 mapped
 		 */
 
-		newsk = tcp_v4_syn_recv_sock(sk, skb, req, dst);
+		newsk = tcp_v4_syn_recv_sock(sk, skb, req, dst,
+					     req_unhash, own_req);
 
 		if (!newsk)
 			return NULL;
@@ -1145,7 +1148,7 @@ static struct sock *tcp_v6_syn_recv_sock(const struct sock *sk, struct sk_buff *
 		tcp_done(newsk);
 		goto out;
 	}
-	__inet_hash(newsk, NULL);
+	*own_req = inet_ehash_nolisten(newsk, req_to_sk(req_unhash));
 
 	return newsk;
 
-- 
cgit v1.2.3