[mirror_ubuntu-artful-kernel.git] / net / ipv4 / tcp_fastopen.c

#include <linux/err.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/list.h>
#include <linux/tcp.h>
#include <linux/rcupdate.h>
#include <linux/rculist.h>
#include <net/inetpeer.h>
#include <net/tcp.h>

int sysctl_tcp_fastopen __read_mostly = TFO_CLIENT_ENABLE;

struct tcp_fastopen_context __rcu *tcp_fastopen_ctx;

static DEFINE_SPINLOCK(tcp_fastopen_ctx_lock);

void tcp_fastopen_init_key_once(bool publish)
{
	static u8 key[TCP_FASTOPEN_KEY_LENGTH];

	/* tcp_fastopen_reset_cipher publishes the new context
	 * atomically, so we allow this race happening here.
	 *
	 * All call sites of tcp_fastopen_cookie_gen also check
	 * for a valid cookie, so this is an acceptable risk.
	 */
	if (net_get_random_once(key, sizeof(key)) && publish)
		tcp_fastopen_reset_cipher(key, sizeof(key));
}

static void tcp_fastopen_ctx_free(struct rcu_head *head)
{
	struct tcp_fastopen_context *ctx =
	    container_of(head, struct tcp_fastopen_context, rcu);
	crypto_free_cipher(ctx->tfm);
	kfree(ctx);
}

int tcp_fastopen_reset_cipher(void *key, unsigned int len)
{
	int err;
	struct tcp_fastopen_context *ctx, *octx;

	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
	if (!ctx)
		return -ENOMEM;
	ctx->tfm = crypto_alloc_cipher("aes", 0, 0);

	if (IS_ERR(ctx->tfm)) {
		err = PTR_ERR(ctx->tfm);
error:		kfree(ctx);
		pr_err("TCP: TFO aes cipher alloc error: %d\n", err);
		return err;
	}
	err = crypto_cipher_setkey(ctx->tfm, key, len);
	if (err) {
		pr_err("TCP: TFO cipher key error: %d\n", err);
		crypto_free_cipher(ctx->tfm);
		goto error;
	}
	memcpy(ctx->key, key, len);

	spin_lock(&tcp_fastopen_ctx_lock);

	octx = rcu_dereference_protected(tcp_fastopen_ctx,
				lockdep_is_held(&tcp_fastopen_ctx_lock));
	rcu_assign_pointer(tcp_fastopen_ctx, ctx);
	spin_unlock(&tcp_fastopen_ctx_lock);

	if (octx)
		call_rcu(&octx->rcu, tcp_fastopen_ctx_free);
	return err;
}

static bool __tcp_fastopen_cookie_gen(const void *path,
				      struct tcp_fastopen_cookie *foc)
{
	struct tcp_fastopen_context *ctx;
	bool ok = false;

	rcu_read_lock();
	ctx = rcu_dereference(tcp_fastopen_ctx);
	if (ctx) {
		crypto_cipher_encrypt_one(ctx->tfm, foc->val, path);
		foc->len = TCP_FASTOPEN_COOKIE_SIZE;
		ok = true;
	}
	rcu_read_unlock();
	return ok;
}

/* Generate the fastopen cookie by doing aes128 encryption on both
 * the source and destination addresses. Pad 0s for IPv4 or IPv4-mapped-IPv6
 * addresses. For the longer IPv6 addresses use CBC-MAC.
 *
 * XXX (TFO) - refactor when TCP_FASTOPEN_COOKIE_SIZE != AES_BLOCK_SIZE.
 */
static bool tcp_fastopen_cookie_gen(struct request_sock *req,
				    struct sk_buff *syn,
				    struct tcp_fastopen_cookie *foc)
{
	if (req->rsk_ops->family == AF_INET) {
		const struct iphdr *iph = ip_hdr(syn);

		__be32 path[4] = { iph->saddr, iph->daddr, 0, 0 };
		return __tcp_fastopen_cookie_gen(path, foc);
	}

#if IS_ENABLED(CONFIG_IPV6)
	if (req->rsk_ops->family == AF_INET6) {
		const struct ipv6hdr *ip6h = ipv6_hdr(syn);
		struct tcp_fastopen_cookie tmp;

		if (__tcp_fastopen_cookie_gen(&ip6h->saddr, &tmp)) {
			struct in6_addr *buf = (struct in6_addr *) tmp.val;
			int i;

			for (i = 0; i < 4; i++)
				buf->s6_addr32[i] ^= ip6h->daddr.s6_addr32[i];
			return __tcp_fastopen_cookie_gen(buf, foc);
		}
	}
#endif
	return false;
}

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;
	struct sock *child;
	u32 end_seq;

	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);
	if (!child)
		return NULL;

	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
	 * only out of the bits carried in the SYN packet.
	 */
	tp = tcp_sk(child);

	tp->fastopen_rsk = req;
	tcp_rsk(req)->tfo_listener = true;

	/* RFC1323: The window in SYN & SYN/ACK segments is never
	 * scaled. So correct it appropriately.
	 */
	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
	 * 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);
	/* 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);
	tcp_init_congestion_control(child);
	tcp_mtup_init(child);
	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.
	 *
	 * XXX (TFO) - we honor a zero-payload TFO request for now,
	 * (any reason not to?) but no need to queue the skb since
	 * there is no data. How about SYN+FIN?
	 */
	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);

		if (likely(skb2)) {
			skb_dst_drop(skb2);
			__skb_pull(skb2, tcp_hdrlen(skb));
			skb_set_owner_r(skb2, child);
			__skb_queue_tail(&child->sk_receive_queue, skb2);
			tp->syn_data_acked = 1;

			/* u64_stats_update_begin(&tp->syncp) not needed here,
			 * as we certainly are not changing upper 32bit value (0)
			 */
			tp->bytes_received = end_seq - TCP_SKB_CB(skb)->seq - 1;
		} else {
			end_seq = TCP_SKB_CB(skb)->seq + 1;
		}
	}
	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.
	 */
	WARN_ON(!req->sk);
	return child;
}

static bool tcp_fastopen_queue_check(struct sock *sk)
{
	struct fastopen_queue *fastopenq;

	/* Make sure the listener has enabled fastopen, and we don't
	 * exceed the max # of pending TFO requests allowed before trying
	 * to validating the cookie in order to avoid burning CPU cycles
	 * unnecessarily.
	 *
	 * XXX (TFO) - The implication of checking the max_qlen before
	 * processing a cookie request is that clients can't differentiate
	 * 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->max_qlen == 0)
		return false;

	if (fastopenq->qlen >= fastopenq->max_qlen) {
		struct request_sock *req1;
		spin_lock(&fastopenq->lock);
		req1 = fastopenq->rskq_rst_head;
		if (!req1 || time_after(req1->rsk_timer.expires, jiffies)) {
			spin_unlock(&fastopenq->lock);
			NET_INC_STATS_BH(sock_net(sk),
					 LINUX_MIB_TCPFASTOPENLISTENOVERFLOW);
			return false;
		}
		fastopenq->rskq_rst_head = req1->dl_next;
		fastopenq->qlen--;
		spin_unlock(&fastopenq->lock);
		reqsk_put(req1);
	}
	return true;
}

/* Returns true if we should perform Fast Open on the SYN. The cookie (foc)
 * may be updated and return the client in the SYN-ACK later. E.g., Fast Open
 * cookie request (foc->len == 0).
 */
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);

	if (!((sysctl_tcp_fastopen & TFO_SERVER_ENABLE) &&
	      (syn_data || foc->len >= 0) &&
	      tcp_fastopen_queue_check(sk))) {
		foc->len = -1;
		return NULL;
	}

	if (syn_data && (sysctl_tcp_fastopen & TFO_SERVER_COOKIE_NOT_REQD))
		goto fastopen;

	if (foc->len >= 0 &&  /* Client presents or requests a cookie */
	    tcp_fastopen_cookie_gen(req, skb, &valid_foc) &&
	    foc->len == TCP_FASTOPEN_COOKIE_SIZE &&
	    foc->len == valid_foc.len &&
	    !memcmp(foc->val, valid_foc.val, foc->len)) {
		/* Cookie is valid. Create a (full) child socket to accept
		 * the data in SYN before returning a SYN-ACK to ack the
		 * data. If we fail to create the socket, fall back and
		 * ack the ISN only but includes the same cookie.
		 *
		 * Note: Data-less SYN with valid cookie is allowed to send
		 * data in SYN_RECV state.
		 */
fastopen:
		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 child;
		}
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENPASSIVEFAIL);
	} else if (foc->len > 0) /* Client presents an invalid cookie */
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENPASSIVEFAIL);

	valid_foc.exp = foc->exp;
	*foc = valid_foc;
	return NULL;
}
Commit	Line	Data
10467163	1	#include <linux/err.h>
2100c8d2 YC	2	#include <linux/init.h>
2100c8d2 YC	3	#include <linux/kernel.h>
10467163 JC	4	#include <linux/list.h>
	5	#include <linux/tcp.h>
	6	#include <linux/rcupdate.h>
	7	#include <linux/rculist.h>
	8	#include <net/inetpeer.h>
	9	#include <net/tcp.h>
2100c8d2	10
0d41cca4	11	int sysctl_tcp_fastopen __read_mostly = TFO_CLIENT_ENABLE;
10467163 JC	12
	13	struct tcp_fastopen_context __rcu *tcp_fastopen_ctx;
	14
	15	static DEFINE_SPINLOCK(tcp_fastopen_ctx_lock);
	16
222e83d2 HFS	17	void tcp_fastopen_init_key_once(bool publish)
	18	{
	19	static u8 key[TCP_FASTOPEN_KEY_LENGTH];
	20
	21	/* tcp_fastopen_reset_cipher publishes the new context
	22	* atomically, so we allow this race happening here.
	23	*
	24	* All call sites of tcp_fastopen_cookie_gen also check
	25	* for a valid cookie, so this is an acceptable risk.
	26	*/
	27	if (net_get_random_once(key, sizeof(key)) && publish)
	28	tcp_fastopen_reset_cipher(key, sizeof(key));
	29	}
	30
10467163 JC	31	static void tcp_fastopen_ctx_free(struct rcu_head *head)
	32	{
	33	struct tcp_fastopen_context *ctx =
	34	container_of(head, struct tcp_fastopen_context, rcu);
	35	crypto_free_cipher(ctx->tfm);
	36	kfree(ctx);
	37	}
	38
	39	int tcp_fastopen_reset_cipher(void *key, unsigned int len)
	40	{
	41	int err;
	42	struct tcp_fastopen_context ctx, octx;
	43
	44	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
	45	if (!ctx)
	46	return -ENOMEM;
	47	ctx->tfm = crypto_alloc_cipher("aes", 0, 0);
	48
	49	if (IS_ERR(ctx->tfm)) {
	50	err = PTR_ERR(ctx->tfm);
	51	error: kfree(ctx);
	52	pr_err("TCP: TFO aes cipher alloc error: %d\n", err);
	53	return err;
	54	}
	55	err = crypto_cipher_setkey(ctx->tfm, key, len);
	56	if (err) {
	57	pr_err("TCP: TFO cipher key error: %d\n", err);
	58	crypto_free_cipher(ctx->tfm);
	59	goto error;
	60	}
	61	memcpy(ctx->key, key, len);
	62
	63	spin_lock(&tcp_fastopen_ctx_lock);
	64
	65	octx = rcu_dereference_protected(tcp_fastopen_ctx,
	66	lockdep_is_held(&tcp_fastopen_ctx_lock));
	67	rcu_assign_pointer(tcp_fastopen_ctx, ctx);
	68	spin_unlock(&tcp_fastopen_ctx_lock);
	69
	70	if (octx)
	71	call_rcu(&octx->rcu, tcp_fastopen_ctx_free);
	72	return err;
	73	}
	74
3a19ce0e DL	75	static bool __tcp_fastopen_cookie_gen(const void *path,
3a19ce0e DL	76	struct tcp_fastopen_cookie *foc)
10467163	77	{
10467163	78	struct tcp_fastopen_context *ctx;
3a19ce0e	79	bool ok = false;
10467163 JC	80
	81	rcu_read_lock();
	82	ctx = rcu_dereference(tcp_fastopen_ctx);
	83	if (ctx) {
3a19ce0e	84	crypto_cipher_encrypt_one(ctx->tfm, foc->val, path);
10467163	85	foc->len = TCP_FASTOPEN_COOKIE_SIZE;
3a19ce0e	86	ok = true;
10467163 JC	87	}
10467163 JC	88	rcu_read_unlock();
3a19ce0e DL	89	return ok;
	90	}
	91
	92	/* Generate the fastopen cookie by doing aes128 encryption on both
	93	* the source and destination addresses. Pad 0s for IPv4 or IPv4-mapped-IPv6
	94	* addresses. For the longer IPv6 addresses use CBC-MAC.
	95	*
	96	* XXX (TFO) - refactor when TCP_FASTOPEN_COOKIE_SIZE != AES_BLOCK_SIZE.
	97	*/
	98	static bool tcp_fastopen_cookie_gen(struct request_sock *req,
	99	struct sk_buff *syn,
	100	struct tcp_fastopen_cookie *foc)
	101	{
	102	if (req->rsk_ops->family == AF_INET) {
	103	const struct iphdr *iph = ip_hdr(syn);
	104
	105	__be32 path[4] = { iph->saddr, iph->daddr, 0, 0 };
	106	return __tcp_fastopen_cookie_gen(path, foc);
	107	}
	108
	109	#if IS_ENABLED(CONFIG_IPV6)
	110	if (req->rsk_ops->family == AF_INET6) {
	111	const struct ipv6hdr *ip6h = ipv6_hdr(syn);
	112	struct tcp_fastopen_cookie tmp;
	113
	114	if (__tcp_fastopen_cookie_gen(&ip6h->saddr, &tmp)) {
	115	struct in6_addr buf = (struct in6_addr ) tmp.val;
41c91996	116	int i;
3a19ce0e DL	117
	118	for (i = 0; i < 4; i++)
	119	buf->s6_addr32[i] ^= ip6h->daddr.s6_addr32[i];
	120	return __tcp_fastopen_cookie_gen(buf, foc);
	121	}
	122	}
	123	#endif
	124	return false;
10467163	125	}
5b7ed089	126
7c85af88 ED	127	static struct sock tcp_fastopen_create_child(struct sock sk,
	128	struct sk_buff *skb,
	129	struct dst_entry *dst,
	130	struct request_sock *req)
5b7ed089	131	{
17846376	132	struct tcp_sock *tp;
5b7ed089	133	struct request_sock_queue *queue = &inet_csk(sk)->icsk_accept_queue;
5b7ed089	134	struct sock *child;
ba34e6d9	135	u32 end_seq;
5b7ed089 YC	136
	137	req->num_retrans = 0;
	138	req->num_timeout = 0;
	139	req->sk = NULL;
	140
	141	child = inet_csk(sk)->icsk_af_ops->syn_recv_sock(sk, skb, req, NULL);
51456b29	142	if (!child)
7c85af88	143	return NULL;
5b7ed089	144
0536fcc0 ED	145	spin_lock(&queue->fastopenq.lock);
	146	queue->fastopenq.qlen++;
	147	spin_unlock(&queue->fastopenq.lock);
5b7ed089 YC	148
	149	/* Initialize the child socket. Have to fix some values to take
	150	* into account the child is a Fast Open socket and is created
	151	* only out of the bits carried in the SYN packet.
	152	*/
	153	tp = tcp_sk(child);
	154
	155	tp->fastopen_rsk = req;
9439ce00	156	tcp_rsk(req)->tfo_listener = true;
5b7ed089 YC	157
	158	/* RFC1323: The window in SYN & SYN/ACK segments is never
	159	* scaled. So correct it appropriately.
	160	*/
	161	tp->snd_wnd = ntohs(tcp_hdr(skb)->window);
	162
	163	/* Activate the retrans timer so that SYNACK can be retransmitted.
	164	* The request socket is not added to the SYN table of the parent
	165	* because it's been added to the accept queue directly.
	166	*/
	167	inet_csk_reset_xmit_timer(child, ICSK_TIME_RETRANS,
	168	TCP_TIMEOUT_INIT, TCP_RTO_MAX);
	169
0470c8ca	170	atomic_set(&req->rsk_refcnt, 1);
5b7ed089 YC	171	/* Add the child socket directly into the accept queue */
	172	inet_csk_reqsk_queue_add(sk, req, child);
	173
	174	/* Now finish processing the fastopen child socket. */
	175	inet_csk(child)->icsk_af_ops->rebuild_header(child);
	176	tcp_init_congestion_control(child);
	177	tcp_mtup_init(child);
	178	tcp_init_metrics(child);
	179	tcp_init_buffer_space(child);
	180
	181	/* Queue the data carried in the SYN packet. We need to first
	182	* bump skb's refcnt because the caller will attempt to free it.
ba34e6d9 ED	183	* Note that IPv6 might also have used skb_get() trick
	184	* in tcp_v6_conn_request() to keep this SYN around (treq->pktopts)
	185	* So we need to eventually get a clone of the packet,
	186	* before inserting it in sk_receive_queue.
5b7ed089	187	*
843f4a55 YC	188	* XXX (TFO) - we honor a zero-payload TFO request for now,
	189	* (any reason not to?) but no need to queue the skb since
	190	* there is no data. How about SYN+FIN?
5b7ed089	191	*/
ba34e6d9 ED	192	end_seq = TCP_SKB_CB(skb)->end_seq;
	193	if (end_seq != TCP_SKB_CB(skb)->seq + 1) {
	194	struct sk_buff *skb2;
	195
	196	if (unlikely(skb_shared(skb)))
	197	skb2 = skb_clone(skb, GFP_ATOMIC);
	198	else
	199	skb2 = skb_get(skb);
	200
	201	if (likely(skb2)) {
	202	skb_dst_drop(skb2);
	203	__skb_pull(skb2, tcp_hdrlen(skb));
	204	skb_set_owner_r(skb2, child);
	205	__skb_queue_tail(&child->sk_receive_queue, skb2);
	206	tp->syn_data_acked = 1;
d654976c ED	207
	208	/* u64_stats_update_begin(&tp->syncp) not needed here,
	209	* as we certainly are not changing upper 32bit value (0)
	210	*/
bdd1f9ed	211	tp->bytes_received = end_seq - TCP_SKB_CB(skb)->seq - 1;
ba34e6d9 ED	212	} else {
	213	end_seq = TCP_SKB_CB(skb)->seq + 1;
	214	}
5b7ed089	215	}
ba34e6d9	216	tcp_rsk(req)->rcv_nxt = tp->rcv_nxt = end_seq;
5b7ed089 YC	217	sk->sk_data_ready(sk);
5b7ed089 YC	218	bh_unlock_sock(child);
7c85af88 ED	219	/* Note: sock_put(child) will be done by tcp_conn_request()
	220	* after SYNACK packet is sent.
	221	*/
51456b29	222	WARN_ON(!req->sk);
7c85af88	223	return child;
5b7ed089	224	}
5b7ed089 YC	225
	226	static bool tcp_fastopen_queue_check(struct sock *sk)
	227	{
	228	struct fastopen_queue *fastopenq;
	229
	230	/* Make sure the listener has enabled fastopen, and we don't
	231	* exceed the max # of pending TFO requests allowed before trying
	232	* to validating the cookie in order to avoid burning CPU cycles
	233	* unnecessarily.
	234	*
	235	* XXX (TFO) - The implication of checking the max_qlen before
	236	* processing a cookie request is that clients can't differentiate
	237	* between qlen overflow causing Fast Open to be disabled
	238	* temporarily vs a server not supporting Fast Open at all.
	239	*/
0536fcc0 ED	240	fastopenq = &inet_csk(sk)->icsk_accept_queue.fastopenq;
0536fcc0 ED	241	if (fastopenq->max_qlen == 0)
5b7ed089 YC	242	return false;
	243
	244	if (fastopenq->qlen >= fastopenq->max_qlen) {
	245	struct request_sock *req1;
	246	spin_lock(&fastopenq->lock);
	247	req1 = fastopenq->rskq_rst_head;
fa76ce73	248	if (!req1 \|\| time_after(req1->rsk_timer.expires, jiffies)) {
5b7ed089 YC	249	spin_unlock(&fastopenq->lock);
	250	NET_INC_STATS_BH(sock_net(sk),
	251	LINUX_MIB_TCPFASTOPENLISTENOVERFLOW);
	252	return false;
	253	}
	254	fastopenq->rskq_rst_head = req1->dl_next;
	255	fastopenq->qlen--;
	256	spin_unlock(&fastopenq->lock);
13854e5a	257	reqsk_put(req1);
5b7ed089 YC	258	}
	259	return true;
	260	}
	261
89278c9d YC	262	/* Returns true if we should perform Fast Open on the SYN. The cookie (foc)
	263	* may be updated and return the client in the SYN-ACK later. E.g., Fast Open
	264	* cookie request (foc->len == 0).
	265	*/
7c85af88 ED	266	struct sock tcp_try_fastopen(struct sock sk, struct sk_buff *skb,
	267	struct request_sock *req,
	268	struct tcp_fastopen_cookie *foc,
	269	struct dst_entry *dst)
5b7ed089	270	{
89278c9d YC	271	struct tcp_fastopen_cookie valid_foc = { .len = -1 };
89278c9d YC	272	bool syn_data = TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq + 1;
7c85af88	273	struct sock *child;
5b7ed089	274
531c94a9 YC	275	if (foc->len == 0) /* Client requests a cookie */
	276	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENCOOKIEREQD);
	277
89278c9d YC	278	if (!((sysctl_tcp_fastopen & TFO_SERVER_ENABLE) &&
	279	(syn_data \|\| foc->len >= 0) &&
	280	tcp_fastopen_queue_check(sk))) {
	281	foc->len = -1;
7c85af88	282	return NULL;
5b7ed089 YC	283	}
5b7ed089 YC	284
89278c9d YC	285	if (syn_data && (sysctl_tcp_fastopen & TFO_SERVER_COOKIE_NOT_REQD))
	286	goto fastopen;
	287
531c94a9 YC	288	if (foc->len >= 0 && /* Client presents or requests a cookie */
531c94a9 YC	289	tcp_fastopen_cookie_gen(req, skb, &valid_foc) &&
3a19ce0e	290	foc->len == TCP_FASTOPEN_COOKIE_SIZE &&
89278c9d YC	291	foc->len == valid_foc.len &&
89278c9d YC	292	!memcmp(foc->val, valid_foc.val, foc->len)) {
843f4a55 YC	293	/* Cookie is valid. Create a (full) child socket to accept
	294	* the data in SYN before returning a SYN-ACK to ack the
	295	* data. If we fail to create the socket, fall back and
	296	* ack the ISN only but includes the same cookie.
	297	*
	298	* Note: Data-less SYN with valid cookie is allowed to send
	299	* data in SYN_RECV state.
	300	*/
89278c9d	301	fastopen:
7c85af88 ED	302	child = tcp_fastopen_create_child(sk, skb, dst, req);
7c85af88 ED	303	if (child) {
843f4a55 YC	304	foc->len = -1;
	305	NET_INC_STATS_BH(sock_net(sk),
	306	LINUX_MIB_TCPFASTOPENPASSIVE);
7c85af88	307	return child;
843f4a55	308	}
531c94a9 YC	309	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENPASSIVEFAIL);
	310	} else if (foc->len > 0) /* Client presents an invalid cookie */
	311	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFASTOPENPASSIVEFAIL);
89278c9d	312
7f9b838b	313	valid_foc.exp = foc->exp;
89278c9d	314	*foc = valid_foc;
7c85af88	315	return NULL;
5b7ed089	316	}