xref: /kernel/linux/linux-5.10/net/dccp/output.c (revision 8c2ecf20)
1// SPDX-License-Identifier: GPL-2.0-or-later
2/*
3 *  net/dccp/output.c
4 *
5 *  An implementation of the DCCP protocol
6 *  Arnaldo Carvalho de Melo <acme@conectiva.com.br>
7 */
8
9#include <linux/dccp.h>
10#include <linux/kernel.h>
11#include <linux/skbuff.h>
12#include <linux/slab.h>
13#include <linux/sched/signal.h>
14
15#include <net/inet_sock.h>
16#include <net/sock.h>
17
18#include "ackvec.h"
19#include "ccid.h"
20#include "dccp.h"
21
22static inline void dccp_event_ack_sent(struct sock *sk)
23{
24	inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK);
25}
26
27/* enqueue @skb on sk_send_head for retransmission, return clone to send now */
28static struct sk_buff *dccp_skb_entail(struct sock *sk, struct sk_buff *skb)
29{
30	skb_set_owner_w(skb, sk);
31	WARN_ON(sk->sk_send_head);
32	sk->sk_send_head = skb;
33	return skb_clone(sk->sk_send_head, gfp_any());
34}
35
36/*
37 * All SKB's seen here are completely headerless. It is our
38 * job to build the DCCP header, and pass the packet down to
39 * IP so it can do the same plus pass the packet off to the
40 * device.
41 */
42static int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb)
43{
44	if (likely(skb != NULL)) {
45		struct inet_sock *inet = inet_sk(sk);
46		const struct inet_connection_sock *icsk = inet_csk(sk);
47		struct dccp_sock *dp = dccp_sk(sk);
48		struct dccp_skb_cb *dcb = DCCP_SKB_CB(skb);
49		struct dccp_hdr *dh;
50		/* XXX For now we're using only 48 bits sequence numbers */
51		const u32 dccp_header_size = sizeof(*dh) +
52					     sizeof(struct dccp_hdr_ext) +
53					  dccp_packet_hdr_len(dcb->dccpd_type);
54		int err, set_ack = 1;
55		u64 ackno = dp->dccps_gsr;
56		/*
57		 * Increment GSS here already in case the option code needs it.
58		 * Update GSS for real only if option processing below succeeds.
59		 */
60		dcb->dccpd_seq = ADD48(dp->dccps_gss, 1);
61
62		switch (dcb->dccpd_type) {
63		case DCCP_PKT_DATA:
64			set_ack = 0;
65			fallthrough;
66		case DCCP_PKT_DATAACK:
67		case DCCP_PKT_RESET:
68			break;
69
70		case DCCP_PKT_REQUEST:
71			set_ack = 0;
72			/* Use ISS on the first (non-retransmitted) Request. */
73			if (icsk->icsk_retransmits == 0)
74				dcb->dccpd_seq = dp->dccps_iss;
75			fallthrough;
76
77		case DCCP_PKT_SYNC:
78		case DCCP_PKT_SYNCACK:
79			ackno = dcb->dccpd_ack_seq;
80			fallthrough;
81		default:
82			/*
83			 * Set owner/destructor: some skbs are allocated via
84			 * alloc_skb (e.g. when retransmission may happen).
85			 * Only Data, DataAck, and Reset packets should come
86			 * through here with skb->sk set.
87			 */
88			WARN_ON(skb->sk);
89			skb_set_owner_w(skb, sk);
90			break;
91		}
92
93		if (dccp_insert_options(sk, skb)) {
94			kfree_skb(skb);
95			return -EPROTO;
96		}
97
98
99		/* Build DCCP header and checksum it. */
100		dh = dccp_zeroed_hdr(skb, dccp_header_size);
101		dh->dccph_type	= dcb->dccpd_type;
102		dh->dccph_sport	= inet->inet_sport;
103		dh->dccph_dport	= inet->inet_dport;
104		dh->dccph_doff	= (dccp_header_size + dcb->dccpd_opt_len) / 4;
105		dh->dccph_ccval	= dcb->dccpd_ccval;
106		dh->dccph_cscov = dp->dccps_pcslen;
107		/* XXX For now we're using only 48 bits sequence numbers */
108		dh->dccph_x	= 1;
109
110		dccp_update_gss(sk, dcb->dccpd_seq);
111		dccp_hdr_set_seq(dh, dp->dccps_gss);
112		if (set_ack)
113			dccp_hdr_set_ack(dccp_hdr_ack_bits(skb), ackno);
114
115		switch (dcb->dccpd_type) {
116		case DCCP_PKT_REQUEST:
117			dccp_hdr_request(skb)->dccph_req_service =
118							dp->dccps_service;
119			/*
120			 * Limit Ack window to ISS <= P.ackno <= GSS, so that
121			 * only Responses to Requests we sent are considered.
122			 */
123			dp->dccps_awl = dp->dccps_iss;
124			break;
125		case DCCP_PKT_RESET:
126			dccp_hdr_reset(skb)->dccph_reset_code =
127							dcb->dccpd_reset_code;
128			break;
129		}
130
131		icsk->icsk_af_ops->send_check(sk, skb);
132
133		if (set_ack)
134			dccp_event_ack_sent(sk);
135
136		DCCP_INC_STATS(DCCP_MIB_OUTSEGS);
137
138		err = icsk->icsk_af_ops->queue_xmit(sk, skb, &inet->cork.fl);
139		return net_xmit_eval(err);
140	}
141	return -ENOBUFS;
142}
143
144/**
145 * dccp_determine_ccmps  -  Find out about CCID-specific packet-size limits
146 * We only consider the HC-sender CCID for setting the CCMPS (RFC 4340, 14.),
147 * since the RX CCID is restricted to feedback packets (Acks), which are small
148 * in comparison with the data traffic. A value of 0 means "no current CCMPS".
149 */
150static u32 dccp_determine_ccmps(const struct dccp_sock *dp)
151{
152	const struct ccid *tx_ccid = dp->dccps_hc_tx_ccid;
153
154	if (tx_ccid == NULL || tx_ccid->ccid_ops == NULL)
155		return 0;
156	return tx_ccid->ccid_ops->ccid_ccmps;
157}
158
159unsigned int dccp_sync_mss(struct sock *sk, u32 pmtu)
160{
161	struct inet_connection_sock *icsk = inet_csk(sk);
162	struct dccp_sock *dp = dccp_sk(sk);
163	u32 ccmps = dccp_determine_ccmps(dp);
164	u32 cur_mps = ccmps ? min(pmtu, ccmps) : pmtu;
165
166	/* Account for header lengths and IPv4/v6 option overhead */
167	cur_mps -= (icsk->icsk_af_ops->net_header_len + icsk->icsk_ext_hdr_len +
168		    sizeof(struct dccp_hdr) + sizeof(struct dccp_hdr_ext));
169
170	/*
171	 * Leave enough headroom for common DCCP header options.
172	 * This only considers options which may appear on DCCP-Data packets, as
173	 * per table 3 in RFC 4340, 5.8. When running out of space for other
174	 * options (eg. Ack Vector which can take up to 255 bytes), it is better
175	 * to schedule a separate Ack. Thus we leave headroom for the following:
176	 *  - 1 byte for Slow Receiver (11.6)
177	 *  - 6 bytes for Timestamp (13.1)
178	 *  - 10 bytes for Timestamp Echo (13.3)
179	 *  - 8 bytes for NDP count (7.7, when activated)
180	 *  - 6 bytes for Data Checksum (9.3)
181	 *  - %DCCPAV_MIN_OPTLEN bytes for Ack Vector size (11.4, when enabled)
182	 */
183	cur_mps -= roundup(1 + 6 + 10 + dp->dccps_send_ndp_count * 8 + 6 +
184			   (dp->dccps_hc_rx_ackvec ? DCCPAV_MIN_OPTLEN : 0), 4);
185
186	/* And store cached results */
187	icsk->icsk_pmtu_cookie = pmtu;
188	WRITE_ONCE(dp->dccps_mss_cache, cur_mps);
189
190	return cur_mps;
191}
192
193EXPORT_SYMBOL_GPL(dccp_sync_mss);
194
195void dccp_write_space(struct sock *sk)
196{
197	struct socket_wq *wq;
198
199	rcu_read_lock();
200	wq = rcu_dereference(sk->sk_wq);
201	if (skwq_has_sleeper(wq))
202		wake_up_interruptible(&wq->wait);
203	/* Should agree with poll, otherwise some programs break */
204	if (sock_writeable(sk))
205		sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
206
207	rcu_read_unlock();
208}
209
210/**
211 * dccp_wait_for_ccid  -  Await CCID send permission
212 * @sk:    socket to wait for
213 * @delay: timeout in jiffies
214 *
215 * This is used by CCIDs which need to delay the send time in process context.
216 */
217static int dccp_wait_for_ccid(struct sock *sk, unsigned long delay)
218{
219	DEFINE_WAIT(wait);
220	long remaining;
221
222	prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
223	sk->sk_write_pending++;
224	release_sock(sk);
225
226	remaining = schedule_timeout(delay);
227
228	lock_sock(sk);
229	sk->sk_write_pending--;
230	finish_wait(sk_sleep(sk), &wait);
231
232	if (signal_pending(current) || sk->sk_err)
233		return -1;
234	return remaining;
235}
236
237/**
238 * dccp_xmit_packet  -  Send data packet under control of CCID
239 * Transmits next-queued payload and informs CCID to account for the packet.
240 */
241static void dccp_xmit_packet(struct sock *sk)
242{
243	int err, len;
244	struct dccp_sock *dp = dccp_sk(sk);
245	struct sk_buff *skb = dccp_qpolicy_pop(sk);
246
247	if (unlikely(skb == NULL))
248		return;
249	len = skb->len;
250
251	if (sk->sk_state == DCCP_PARTOPEN) {
252		const u32 cur_mps = dp->dccps_mss_cache - DCCP_FEATNEG_OVERHEAD;
253		/*
254		 * See 8.1.5 - Handshake Completion.
255		 *
256		 * For robustness we resend Confirm options until the client has
257		 * entered OPEN. During the initial feature negotiation, the MPS
258		 * is smaller than usual, reduced by the Change/Confirm options.
259		 */
260		if (!list_empty(&dp->dccps_featneg) && len > cur_mps) {
261			DCCP_WARN("Payload too large (%d) for featneg.\n", len);
262			dccp_send_ack(sk);
263			dccp_feat_list_purge(&dp->dccps_featneg);
264		}
265
266		inet_csk_schedule_ack(sk);
267		inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK,
268					      inet_csk(sk)->icsk_rto,
269					      DCCP_RTO_MAX);
270		DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_DATAACK;
271	} else if (dccp_ack_pending(sk)) {
272		DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_DATAACK;
273	} else {
274		DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_DATA;
275	}
276
277	err = dccp_transmit_skb(sk, skb);
278	if (err)
279		dccp_pr_debug("transmit_skb() returned err=%d\n", err);
280	/*
281	 * Register this one as sent even if an error occurred. To the remote
282	 * end a local packet drop is indistinguishable from network loss, i.e.
283	 * any local drop will eventually be reported via receiver feedback.
284	 */
285	ccid_hc_tx_packet_sent(dp->dccps_hc_tx_ccid, sk, len);
286
287	/*
288	 * If the CCID needs to transfer additional header options out-of-band
289	 * (e.g. Ack Vectors or feature-negotiation options), it activates this
290	 * flag to schedule a Sync. The Sync will automatically incorporate all
291	 * currently pending header options, thus clearing the backlog.
292	 */
293	if (dp->dccps_sync_scheduled)
294		dccp_send_sync(sk, dp->dccps_gsr, DCCP_PKT_SYNC);
295}
296
297/**
298 * dccp_flush_write_queue  -  Drain queue at end of connection
299 * Since dccp_sendmsg queues packets without waiting for them to be sent, it may
300 * happen that the TX queue is not empty at the end of a connection. We give the
301 * HC-sender CCID a grace period of up to @time_budget jiffies. If this function
302 * returns with a non-empty write queue, it will be purged later.
303 */
304void dccp_flush_write_queue(struct sock *sk, long *time_budget)
305{
306	struct dccp_sock *dp = dccp_sk(sk);
307	struct sk_buff *skb;
308	long delay, rc;
309
310	while (*time_budget > 0 && (skb = skb_peek(&sk->sk_write_queue))) {
311		rc = ccid_hc_tx_send_packet(dp->dccps_hc_tx_ccid, sk, skb);
312
313		switch (ccid_packet_dequeue_eval(rc)) {
314		case CCID_PACKET_WILL_DEQUEUE_LATER:
315			/*
316			 * If the CCID determines when to send, the next sending
317			 * time is unknown or the CCID may not even send again
318			 * (e.g. remote host crashes or lost Ack packets).
319			 */
320			DCCP_WARN("CCID did not manage to send all packets\n");
321			return;
322		case CCID_PACKET_DELAY:
323			delay = msecs_to_jiffies(rc);
324			if (delay > *time_budget)
325				return;
326			rc = dccp_wait_for_ccid(sk, delay);
327			if (rc < 0)
328				return;
329			*time_budget -= (delay - rc);
330			/* check again if we can send now */
331			break;
332		case CCID_PACKET_SEND_AT_ONCE:
333			dccp_xmit_packet(sk);
334			break;
335		case CCID_PACKET_ERR:
336			skb_dequeue(&sk->sk_write_queue);
337			kfree_skb(skb);
338			dccp_pr_debug("packet discarded due to err=%ld\n", rc);
339		}
340	}
341}
342
343void dccp_write_xmit(struct sock *sk)
344{
345	struct dccp_sock *dp = dccp_sk(sk);
346	struct sk_buff *skb;
347
348	while ((skb = dccp_qpolicy_top(sk))) {
349		int rc = ccid_hc_tx_send_packet(dp->dccps_hc_tx_ccid, sk, skb);
350
351		switch (ccid_packet_dequeue_eval(rc)) {
352		case CCID_PACKET_WILL_DEQUEUE_LATER:
353			return;
354		case CCID_PACKET_DELAY:
355			sk_reset_timer(sk, &dp->dccps_xmit_timer,
356				       jiffies + msecs_to_jiffies(rc));
357			return;
358		case CCID_PACKET_SEND_AT_ONCE:
359			dccp_xmit_packet(sk);
360			break;
361		case CCID_PACKET_ERR:
362			dccp_qpolicy_drop(sk, skb);
363			dccp_pr_debug("packet discarded due to err=%d\n", rc);
364		}
365	}
366}
367
368/**
369 * dccp_retransmit_skb  -  Retransmit Request, Close, or CloseReq packets
370 * There are only four retransmittable packet types in DCCP:
371 * - Request  in client-REQUEST  state (sec. 8.1.1),
372 * - CloseReq in server-CLOSEREQ state (sec. 8.3),
373 * - Close    in   node-CLOSING  state (sec. 8.3),
374 * - Acks in client-PARTOPEN state (sec. 8.1.5, handled by dccp_delack_timer()).
375 * This function expects sk->sk_send_head to contain the original skb.
376 */
377int dccp_retransmit_skb(struct sock *sk)
378{
379	WARN_ON(sk->sk_send_head == NULL);
380
381	if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk) != 0)
382		return -EHOSTUNREACH; /* Routing failure or similar. */
383
384	/* this count is used to distinguish original and retransmitted skb */
385	inet_csk(sk)->icsk_retransmits++;
386
387	return dccp_transmit_skb(sk, skb_clone(sk->sk_send_head, GFP_ATOMIC));
388}
389
390struct sk_buff *dccp_make_response(const struct sock *sk, struct dst_entry *dst,
391				   struct request_sock *req)
392{
393	struct dccp_hdr *dh;
394	struct dccp_request_sock *dreq;
395	const u32 dccp_header_size = sizeof(struct dccp_hdr) +
396				     sizeof(struct dccp_hdr_ext) +
397				     sizeof(struct dccp_hdr_response);
398	struct sk_buff *skb;
399
400	/* sk is marked const to clearly express we dont hold socket lock.
401	 * sock_wmalloc() will atomically change sk->sk_wmem_alloc,
402	 * it is safe to promote sk to non const.
403	 */
404	skb = sock_wmalloc((struct sock *)sk, MAX_DCCP_HEADER, 1,
405			   GFP_ATOMIC);
406	if (!skb)
407		return NULL;
408
409	skb_reserve(skb, MAX_DCCP_HEADER);
410
411	skb_dst_set(skb, dst_clone(dst));
412
413	dreq = dccp_rsk(req);
414	if (inet_rsk(req)->acked)	/* increase GSS upon retransmission */
415		dccp_inc_seqno(&dreq->dreq_gss);
416	DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_RESPONSE;
417	DCCP_SKB_CB(skb)->dccpd_seq  = dreq->dreq_gss;
418
419	/* Resolve feature dependencies resulting from choice of CCID */
420	if (dccp_feat_server_ccid_dependencies(dreq))
421		goto response_failed;
422
423	if (dccp_insert_options_rsk(dreq, skb))
424		goto response_failed;
425
426	/* Build and checksum header */
427	dh = dccp_zeroed_hdr(skb, dccp_header_size);
428
429	dh->dccph_sport	= htons(inet_rsk(req)->ir_num);
430	dh->dccph_dport	= inet_rsk(req)->ir_rmt_port;
431	dh->dccph_doff	= (dccp_header_size +
432			   DCCP_SKB_CB(skb)->dccpd_opt_len) / 4;
433	dh->dccph_type	= DCCP_PKT_RESPONSE;
434	dh->dccph_x	= 1;
435	dccp_hdr_set_seq(dh, dreq->dreq_gss);
436	dccp_hdr_set_ack(dccp_hdr_ack_bits(skb), dreq->dreq_gsr);
437	dccp_hdr_response(skb)->dccph_resp_service = dreq->dreq_service;
438
439	dccp_csum_outgoing(skb);
440
441	/* We use `acked' to remember that a Response was already sent. */
442	inet_rsk(req)->acked = 1;
443	DCCP_INC_STATS(DCCP_MIB_OUTSEGS);
444	return skb;
445response_failed:
446	kfree_skb(skb);
447	return NULL;
448}
449
450EXPORT_SYMBOL_GPL(dccp_make_response);
451
452/* answer offending packet in @rcv_skb with Reset from control socket @ctl */
453struct sk_buff *dccp_ctl_make_reset(struct sock *sk, struct sk_buff *rcv_skb)
454{
455	struct dccp_hdr *rxdh = dccp_hdr(rcv_skb), *dh;
456	struct dccp_skb_cb *dcb = DCCP_SKB_CB(rcv_skb);
457	const u32 dccp_hdr_reset_len = sizeof(struct dccp_hdr) +
458				       sizeof(struct dccp_hdr_ext) +
459				       sizeof(struct dccp_hdr_reset);
460	struct dccp_hdr_reset *dhr;
461	struct sk_buff *skb;
462
463	skb = alloc_skb(sk->sk_prot->max_header, GFP_ATOMIC);
464	if (skb == NULL)
465		return NULL;
466
467	skb_reserve(skb, sk->sk_prot->max_header);
468
469	/* Swap the send and the receive. */
470	dh = dccp_zeroed_hdr(skb, dccp_hdr_reset_len);
471	dh->dccph_type	= DCCP_PKT_RESET;
472	dh->dccph_sport	= rxdh->dccph_dport;
473	dh->dccph_dport	= rxdh->dccph_sport;
474	dh->dccph_doff	= dccp_hdr_reset_len / 4;
475	dh->dccph_x	= 1;
476
477	dhr = dccp_hdr_reset(skb);
478	dhr->dccph_reset_code = dcb->dccpd_reset_code;
479
480	switch (dcb->dccpd_reset_code) {
481	case DCCP_RESET_CODE_PACKET_ERROR:
482		dhr->dccph_reset_data[0] = rxdh->dccph_type;
483		break;
484	case DCCP_RESET_CODE_OPTION_ERROR:
485	case DCCP_RESET_CODE_MANDATORY_ERROR:
486		memcpy(dhr->dccph_reset_data, dcb->dccpd_reset_data, 3);
487		break;
488	}
489	/*
490	 * From RFC 4340, 8.3.1:
491	 *   If P.ackno exists, set R.seqno := P.ackno + 1.
492	 *   Else set R.seqno := 0.
493	 */
494	if (dcb->dccpd_ack_seq != DCCP_PKT_WITHOUT_ACK_SEQ)
495		dccp_hdr_set_seq(dh, ADD48(dcb->dccpd_ack_seq, 1));
496	dccp_hdr_set_ack(dccp_hdr_ack_bits(skb), dcb->dccpd_seq);
497
498	dccp_csum_outgoing(skb);
499	return skb;
500}
501
502EXPORT_SYMBOL_GPL(dccp_ctl_make_reset);
503
504/* send Reset on established socket, to close or abort the connection */
505int dccp_send_reset(struct sock *sk, enum dccp_reset_codes code)
506{
507	struct sk_buff *skb;
508	/*
509	 * FIXME: what if rebuild_header fails?
510	 * Should we be doing a rebuild_header here?
511	 */
512	int err = inet_csk(sk)->icsk_af_ops->rebuild_header(sk);
513
514	if (err != 0)
515		return err;
516
517	skb = sock_wmalloc(sk, sk->sk_prot->max_header, 1, GFP_ATOMIC);
518	if (skb == NULL)
519		return -ENOBUFS;
520
521	/* Reserve space for headers and prepare control bits. */
522	skb_reserve(skb, sk->sk_prot->max_header);
523	DCCP_SKB_CB(skb)->dccpd_type	   = DCCP_PKT_RESET;
524	DCCP_SKB_CB(skb)->dccpd_reset_code = code;
525
526	return dccp_transmit_skb(sk, skb);
527}
528
529/*
530 * Do all connect socket setups that can be done AF independent.
531 */
532int dccp_connect(struct sock *sk)
533{
534	struct sk_buff *skb;
535	struct dccp_sock *dp = dccp_sk(sk);
536	struct dst_entry *dst = __sk_dst_get(sk);
537	struct inet_connection_sock *icsk = inet_csk(sk);
538
539	sk->sk_err = 0;
540	sock_reset_flag(sk, SOCK_DONE);
541
542	dccp_sync_mss(sk, dst_mtu(dst));
543
544	/* do not connect if feature negotiation setup fails */
545	if (dccp_feat_finalise_settings(dccp_sk(sk)))
546		return -EPROTO;
547
548	/* Initialise GAR as per 8.5; AWL/AWH are set in dccp_transmit_skb() */
549	dp->dccps_gar = dp->dccps_iss;
550
551	skb = alloc_skb(sk->sk_prot->max_header, sk->sk_allocation);
552	if (unlikely(skb == NULL))
553		return -ENOBUFS;
554
555	/* Reserve space for headers. */
556	skb_reserve(skb, sk->sk_prot->max_header);
557
558	DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_REQUEST;
559
560	dccp_transmit_skb(sk, dccp_skb_entail(sk, skb));
561	DCCP_INC_STATS(DCCP_MIB_ACTIVEOPENS);
562
563	/* Timer for repeating the REQUEST until an answer. */
564	icsk->icsk_retransmits = 0;
565	inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
566				  icsk->icsk_rto, DCCP_RTO_MAX);
567	return 0;
568}
569
570EXPORT_SYMBOL_GPL(dccp_connect);
571
572void dccp_send_ack(struct sock *sk)
573{
574	/* If we have been reset, we may not send again. */
575	if (sk->sk_state != DCCP_CLOSED) {
576		struct sk_buff *skb = alloc_skb(sk->sk_prot->max_header,
577						GFP_ATOMIC);
578
579		if (skb == NULL) {
580			inet_csk_schedule_ack(sk);
581			inet_csk(sk)->icsk_ack.ato = TCP_ATO_MIN;
582			inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK,
583						  TCP_DELACK_MAX,
584						  DCCP_RTO_MAX);
585			return;
586		}
587
588		/* Reserve space for headers */
589		skb_reserve(skb, sk->sk_prot->max_header);
590		DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_ACK;
591		dccp_transmit_skb(sk, skb);
592	}
593}
594
595EXPORT_SYMBOL_GPL(dccp_send_ack);
596
597#if 0
598/* FIXME: Is this still necessary (11.3) - currently nowhere used by DCCP. */
599void dccp_send_delayed_ack(struct sock *sk)
600{
601	struct inet_connection_sock *icsk = inet_csk(sk);
602	/*
603	 * FIXME: tune this timer. elapsed time fixes the skew, so no problem
604	 * with using 2s, and active senders also piggyback the ACK into a
605	 * DATAACK packet, so this is really for quiescent senders.
606	 */
607	unsigned long timeout = jiffies + 2 * HZ;
608
609	/* Use new timeout only if there wasn't a older one earlier. */
610	if (icsk->icsk_ack.pending & ICSK_ACK_TIMER) {
611		/* If delack timer was blocked or is about to expire,
612		 * send ACK now.
613		 *
614		 * FIXME: check the "about to expire" part
615		 */
616		if (icsk->icsk_ack.blocked) {
617			dccp_send_ack(sk);
618			return;
619		}
620
621		if (!time_before(timeout, icsk->icsk_ack.timeout))
622			timeout = icsk->icsk_ack.timeout;
623	}
624	icsk->icsk_ack.pending |= ICSK_ACK_SCHED | ICSK_ACK_TIMER;
625	icsk->icsk_ack.timeout = timeout;
626	sk_reset_timer(sk, &icsk->icsk_delack_timer, timeout);
627}
628#endif
629
630void dccp_send_sync(struct sock *sk, const u64 ackno,
631		    const enum dccp_pkt_type pkt_type)
632{
633	/*
634	 * We are not putting this on the write queue, so
635	 * dccp_transmit_skb() will set the ownership to this
636	 * sock.
637	 */
638	struct sk_buff *skb = alloc_skb(sk->sk_prot->max_header, GFP_ATOMIC);
639
640	if (skb == NULL) {
641		/* FIXME: how to make sure the sync is sent? */
642		DCCP_CRIT("could not send %s", dccp_packet_name(pkt_type));
643		return;
644	}
645
646	/* Reserve space for headers and prepare control bits. */
647	skb_reserve(skb, sk->sk_prot->max_header);
648	DCCP_SKB_CB(skb)->dccpd_type = pkt_type;
649	DCCP_SKB_CB(skb)->dccpd_ack_seq = ackno;
650
651	/*
652	 * Clear the flag in case the Sync was scheduled for out-of-band data,
653	 * such as carrying a long Ack Vector.
654	 */
655	dccp_sk(sk)->dccps_sync_scheduled = 0;
656
657	dccp_transmit_skb(sk, skb);
658}
659
660EXPORT_SYMBOL_GPL(dccp_send_sync);
661
662/*
663 * Send a DCCP_PKT_CLOSE/CLOSEREQ. The caller locks the socket for us. This
664 * cannot be allowed to fail queueing a DCCP_PKT_CLOSE/CLOSEREQ frame under
665 * any circumstances.
666 */
667void dccp_send_close(struct sock *sk, const int active)
668{
669	struct dccp_sock *dp = dccp_sk(sk);
670	struct sk_buff *skb;
671	const gfp_t prio = active ? GFP_KERNEL : GFP_ATOMIC;
672
673	skb = alloc_skb(sk->sk_prot->max_header, prio);
674	if (skb == NULL)
675		return;
676
677	/* Reserve space for headers and prepare control bits. */
678	skb_reserve(skb, sk->sk_prot->max_header);
679	if (dp->dccps_role == DCCP_ROLE_SERVER && !dp->dccps_server_timewait)
680		DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_CLOSEREQ;
681	else
682		DCCP_SKB_CB(skb)->dccpd_type = DCCP_PKT_CLOSE;
683
684	if (active) {
685		skb = dccp_skb_entail(sk, skb);
686		/*
687		 * Retransmission timer for active-close: RFC 4340, 8.3 requires
688		 * to retransmit the Close/CloseReq until the CLOSING/CLOSEREQ
689		 * state can be left. The initial timeout is 2 RTTs.
690		 * Since RTT measurement is done by the CCIDs, there is no easy
691		 * way to get an RTT sample. The fallback RTT from RFC 4340, 3.4
692		 * is too low (200ms); we use a high value to avoid unnecessary
693		 * retransmissions when the link RTT is > 0.2 seconds.
694		 * FIXME: Let main module sample RTTs and use that instead.
695		 */
696		inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
697					  DCCP_TIMEOUT_INIT, DCCP_RTO_MAX);
698	}
699	dccp_transmit_skb(sk, skb);
700}
701