• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* Management of Tx window, Tx resend, ACKs and out-of-sequence reception
2  *
3  * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
4  * Written by David Howells (dhowells@redhat.com)
5  *
6  * This program is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU General Public License
8  * as published by the Free Software Foundation; either version
9  * 2 of the License, or (at your option) any later version.
10  */
11 
12 #include <linux/module.h>
13 #include <linux/circ_buf.h>
14 #include <linux/net.h>
15 #include <linux/skbuff.h>
16 #include <linux/slab.h>
17 #include <linux/udp.h>
18 #include <net/sock.h>
19 #include <net/af_rxrpc.h>
20 #include "ar-internal.h"
21 
22 /*
23  * How long to wait before scheduling ACK generation after seeing a
24  * packet with RXRPC_REQUEST_ACK set (in jiffies).
25  */
26 unsigned rxrpc_requested_ack_delay = 1;
27 
28 /*
29  * How long to wait before scheduling an ACK with subtype DELAY (in jiffies).
30  *
31  * We use this when we've received new data packets.  If those packets aren't
32  * all consumed within this time we will send a DELAY ACK if an ACK was not
33  * requested to let the sender know it doesn't need to resend.
34  */
35 unsigned rxrpc_soft_ack_delay = 1 * HZ;
36 
37 /*
38  * How long to wait before scheduling an ACK with subtype IDLE (in jiffies).
39  *
40  * We use this when we've consumed some previously soft-ACK'd packets when
41  * further packets aren't immediately received to decide when to send an IDLE
42  * ACK let the other end know that it can free up its Tx buffer space.
43  */
44 unsigned rxrpc_idle_ack_delay = 0.5 * HZ;
45 
46 /*
47  * Receive window size in packets.  This indicates the maximum number of
48  * unconsumed received packets we're willing to retain in memory.  Once this
49  * limit is hit, we should generate an EXCEEDS_WINDOW ACK and discard further
50  * packets.
51  */
52 unsigned rxrpc_rx_window_size = 32;
53 
54 /*
55  * Maximum Rx MTU size.  This indicates to the sender the size of jumbo packet
56  * made by gluing normal packets together that we're willing to handle.
57  */
58 unsigned rxrpc_rx_mtu = 5692;
59 
60 /*
61  * The maximum number of fragments in a received jumbo packet that we tell the
62  * sender that we're willing to handle.
63  */
64 unsigned rxrpc_rx_jumbo_max = 4;
65 
rxrpc_acks(u8 reason)66 static const char *rxrpc_acks(u8 reason)
67 {
68 	static const char *const str[] = {
69 		"---", "REQ", "DUP", "OOS", "WIN", "MEM", "PNG", "PNR", "DLY",
70 		"IDL", "-?-"
71 	};
72 
73 	if (reason >= ARRAY_SIZE(str))
74 		reason = ARRAY_SIZE(str) - 1;
75 	return str[reason];
76 }
77 
78 static const s8 rxrpc_ack_priority[] = {
79 	[0]				= 0,
80 	[RXRPC_ACK_DELAY]		= 1,
81 	[RXRPC_ACK_REQUESTED]		= 2,
82 	[RXRPC_ACK_IDLE]		= 3,
83 	[RXRPC_ACK_PING_RESPONSE]	= 4,
84 	[RXRPC_ACK_DUPLICATE]		= 5,
85 	[RXRPC_ACK_OUT_OF_SEQUENCE]	= 6,
86 	[RXRPC_ACK_EXCEEDS_WINDOW]	= 7,
87 	[RXRPC_ACK_NOSPACE]		= 8,
88 };
89 
90 /*
91  * propose an ACK be sent
92  */
__rxrpc_propose_ACK(struct rxrpc_call * call,u8 ack_reason,__be32 serial,bool immediate)93 void __rxrpc_propose_ACK(struct rxrpc_call *call, u8 ack_reason,
94 			 __be32 serial, bool immediate)
95 {
96 	unsigned long expiry;
97 	s8 prior = rxrpc_ack_priority[ack_reason];
98 
99 	ASSERTCMP(prior, >, 0);
100 
101 	_enter("{%d},%s,%%%x,%u",
102 	       call->debug_id, rxrpc_acks(ack_reason), ntohl(serial),
103 	       immediate);
104 
105 	if (prior < rxrpc_ack_priority[call->ackr_reason]) {
106 		if (immediate)
107 			goto cancel_timer;
108 		return;
109 	}
110 
111 	/* update DELAY, IDLE, REQUESTED and PING_RESPONSE ACK serial
112 	 * numbers */
113 	if (prior == rxrpc_ack_priority[call->ackr_reason]) {
114 		if (prior <= 4)
115 			call->ackr_serial = serial;
116 		if (immediate)
117 			goto cancel_timer;
118 		return;
119 	}
120 
121 	call->ackr_reason = ack_reason;
122 	call->ackr_serial = serial;
123 
124 	switch (ack_reason) {
125 	case RXRPC_ACK_DELAY:
126 		_debug("run delay timer");
127 		expiry = rxrpc_soft_ack_delay;
128 		goto run_timer;
129 
130 	case RXRPC_ACK_IDLE:
131 		if (!immediate) {
132 			_debug("run defer timer");
133 			expiry = rxrpc_idle_ack_delay;
134 			goto run_timer;
135 		}
136 		goto cancel_timer;
137 
138 	case RXRPC_ACK_REQUESTED:
139 		expiry = rxrpc_requested_ack_delay;
140 		if (!expiry)
141 			goto cancel_timer;
142 		if (!immediate || serial == cpu_to_be32(1)) {
143 			_debug("run defer timer");
144 			goto run_timer;
145 		}
146 
147 	default:
148 		_debug("immediate ACK");
149 		goto cancel_timer;
150 	}
151 
152 run_timer:
153 	expiry += jiffies;
154 	if (!timer_pending(&call->ack_timer) ||
155 	    time_after(call->ack_timer.expires, expiry))
156 		mod_timer(&call->ack_timer, expiry);
157 	return;
158 
159 cancel_timer:
160 	_debug("cancel timer %%%u", ntohl(serial));
161 	try_to_del_timer_sync(&call->ack_timer);
162 	read_lock_bh(&call->state_lock);
163 	if (call->state <= RXRPC_CALL_COMPLETE &&
164 	    !test_and_set_bit(RXRPC_CALL_ACK, &call->events))
165 		rxrpc_queue_call(call);
166 	read_unlock_bh(&call->state_lock);
167 }
168 
169 /*
170  * propose an ACK be sent, locking the call structure
171  */
rxrpc_propose_ACK(struct rxrpc_call * call,u8 ack_reason,__be32 serial,bool immediate)172 void rxrpc_propose_ACK(struct rxrpc_call *call, u8 ack_reason,
173 		       __be32 serial, bool immediate)
174 {
175 	s8 prior = rxrpc_ack_priority[ack_reason];
176 
177 	if (prior > rxrpc_ack_priority[call->ackr_reason]) {
178 		spin_lock_bh(&call->lock);
179 		__rxrpc_propose_ACK(call, ack_reason, serial, immediate);
180 		spin_unlock_bh(&call->lock);
181 	}
182 }
183 
184 /*
185  * set the resend timer
186  */
rxrpc_set_resend(struct rxrpc_call * call,u8 resend,unsigned long resend_at)187 static void rxrpc_set_resend(struct rxrpc_call *call, u8 resend,
188 			     unsigned long resend_at)
189 {
190 	read_lock_bh(&call->state_lock);
191 	if (call->state >= RXRPC_CALL_COMPLETE)
192 		resend = 0;
193 
194 	if (resend & 1) {
195 		_debug("SET RESEND");
196 		set_bit(RXRPC_CALL_RESEND, &call->events);
197 	}
198 
199 	if (resend & 2) {
200 		_debug("MODIFY RESEND TIMER");
201 		set_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
202 		mod_timer(&call->resend_timer, resend_at);
203 	} else {
204 		_debug("KILL RESEND TIMER");
205 		del_timer_sync(&call->resend_timer);
206 		clear_bit(RXRPC_CALL_RESEND_TIMER, &call->events);
207 		clear_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
208 	}
209 	read_unlock_bh(&call->state_lock);
210 }
211 
212 /*
213  * resend packets
214  */
rxrpc_resend(struct rxrpc_call * call)215 static void rxrpc_resend(struct rxrpc_call *call)
216 {
217 	struct rxrpc_skb_priv *sp;
218 	struct rxrpc_header *hdr;
219 	struct sk_buff *txb;
220 	unsigned long *p_txb, resend_at;
221 	int loop, stop;
222 	u8 resend;
223 
224 	_enter("{%d,%d,%d,%d},",
225 	       call->acks_hard, call->acks_unacked,
226 	       atomic_read(&call->sequence),
227 	       CIRC_CNT(call->acks_head, call->acks_tail, call->acks_winsz));
228 
229 	stop = 0;
230 	resend = 0;
231 	resend_at = 0;
232 
233 	for (loop = call->acks_tail;
234 	     loop != call->acks_head || stop;
235 	     loop = (loop + 1) &  (call->acks_winsz - 1)
236 	     ) {
237 		p_txb = call->acks_window + loop;
238 		smp_read_barrier_depends();
239 		if (*p_txb & 1)
240 			continue;
241 
242 		txb = (struct sk_buff *) *p_txb;
243 		sp = rxrpc_skb(txb);
244 
245 		if (sp->need_resend) {
246 			sp->need_resend = false;
247 
248 			/* each Tx packet has a new serial number */
249 			sp->hdr.serial =
250 				htonl(atomic_inc_return(&call->conn->serial));
251 
252 			hdr = (struct rxrpc_header *) txb->head;
253 			hdr->serial = sp->hdr.serial;
254 
255 			_proto("Tx DATA %%%u { #%d }",
256 			       ntohl(sp->hdr.serial), ntohl(sp->hdr.seq));
257 			if (rxrpc_send_packet(call->conn->trans, txb) < 0) {
258 				stop = 0;
259 				sp->resend_at = jiffies + 3;
260 			} else {
261 				sp->resend_at =
262 					jiffies + rxrpc_resend_timeout * HZ;
263 			}
264 		}
265 
266 		if (time_after_eq(jiffies + 1, sp->resend_at)) {
267 			sp->need_resend = true;
268 			resend |= 1;
269 		} else if (resend & 2) {
270 			if (time_before(sp->resend_at, resend_at))
271 				resend_at = sp->resend_at;
272 		} else {
273 			resend_at = sp->resend_at;
274 			resend |= 2;
275 		}
276 	}
277 
278 	rxrpc_set_resend(call, resend, resend_at);
279 	_leave("");
280 }
281 
282 /*
283  * handle resend timer expiry
284  */
rxrpc_resend_timer(struct rxrpc_call * call)285 static void rxrpc_resend_timer(struct rxrpc_call *call)
286 {
287 	struct rxrpc_skb_priv *sp;
288 	struct sk_buff *txb;
289 	unsigned long *p_txb, resend_at;
290 	int loop;
291 	u8 resend;
292 
293 	_enter("%d,%d,%d",
294 	       call->acks_tail, call->acks_unacked, call->acks_head);
295 
296 	if (call->state >= RXRPC_CALL_COMPLETE)
297 		return;
298 
299 	resend = 0;
300 	resend_at = 0;
301 
302 	for (loop = call->acks_unacked;
303 	     loop != call->acks_head;
304 	     loop = (loop + 1) &  (call->acks_winsz - 1)
305 	     ) {
306 		p_txb = call->acks_window + loop;
307 		smp_read_barrier_depends();
308 		txb = (struct sk_buff *) (*p_txb & ~1);
309 		sp = rxrpc_skb(txb);
310 
311 		ASSERT(!(*p_txb & 1));
312 
313 		if (sp->need_resend) {
314 			;
315 		} else if (time_after_eq(jiffies + 1, sp->resend_at)) {
316 			sp->need_resend = true;
317 			resend |= 1;
318 		} else if (resend & 2) {
319 			if (time_before(sp->resend_at, resend_at))
320 				resend_at = sp->resend_at;
321 		} else {
322 			resend_at = sp->resend_at;
323 			resend |= 2;
324 		}
325 	}
326 
327 	rxrpc_set_resend(call, resend, resend_at);
328 	_leave("");
329 }
330 
331 /*
332  * process soft ACKs of our transmitted packets
333  * - these indicate packets the peer has or has not received, but hasn't yet
334  *   given to the consumer, and so can still be discarded and re-requested
335  */
rxrpc_process_soft_ACKs(struct rxrpc_call * call,struct rxrpc_ackpacket * ack,struct sk_buff * skb)336 static int rxrpc_process_soft_ACKs(struct rxrpc_call *call,
337 				   struct rxrpc_ackpacket *ack,
338 				   struct sk_buff *skb)
339 {
340 	struct rxrpc_skb_priv *sp;
341 	struct sk_buff *txb;
342 	unsigned long *p_txb, resend_at;
343 	int loop;
344 	u8 sacks[RXRPC_MAXACKS], resend;
345 
346 	_enter("{%d,%d},{%d},",
347 	       call->acks_hard,
348 	       CIRC_CNT(call->acks_head, call->acks_tail, call->acks_winsz),
349 	       ack->nAcks);
350 
351 	if (skb_copy_bits(skb, 0, sacks, ack->nAcks) < 0)
352 		goto protocol_error;
353 
354 	resend = 0;
355 	resend_at = 0;
356 	for (loop = 0; loop < ack->nAcks; loop++) {
357 		p_txb = call->acks_window;
358 		p_txb += (call->acks_tail + loop) & (call->acks_winsz - 1);
359 		smp_read_barrier_depends();
360 		txb = (struct sk_buff *) (*p_txb & ~1);
361 		sp = rxrpc_skb(txb);
362 
363 		switch (sacks[loop]) {
364 		case RXRPC_ACK_TYPE_ACK:
365 			sp->need_resend = false;
366 			*p_txb |= 1;
367 			break;
368 		case RXRPC_ACK_TYPE_NACK:
369 			sp->need_resend = true;
370 			*p_txb &= ~1;
371 			resend = 1;
372 			break;
373 		default:
374 			_debug("Unsupported ACK type %d", sacks[loop]);
375 			goto protocol_error;
376 		}
377 	}
378 
379 	smp_mb();
380 	call->acks_unacked = (call->acks_tail + loop) & (call->acks_winsz - 1);
381 
382 	/* anything not explicitly ACK'd is implicitly NACK'd, but may just not
383 	 * have been received or processed yet by the far end */
384 	for (loop = call->acks_unacked;
385 	     loop != call->acks_head;
386 	     loop = (loop + 1) &  (call->acks_winsz - 1)
387 	     ) {
388 		p_txb = call->acks_window + loop;
389 		smp_read_barrier_depends();
390 		txb = (struct sk_buff *) (*p_txb & ~1);
391 		sp = rxrpc_skb(txb);
392 
393 		if (*p_txb & 1) {
394 			/* packet must have been discarded */
395 			sp->need_resend = true;
396 			*p_txb &= ~1;
397 			resend |= 1;
398 		} else if (sp->need_resend) {
399 			;
400 		} else if (time_after_eq(jiffies + 1, sp->resend_at)) {
401 			sp->need_resend = true;
402 			resend |= 1;
403 		} else if (resend & 2) {
404 			if (time_before(sp->resend_at, resend_at))
405 				resend_at = sp->resend_at;
406 		} else {
407 			resend_at = sp->resend_at;
408 			resend |= 2;
409 		}
410 	}
411 
412 	rxrpc_set_resend(call, resend, resend_at);
413 	_leave(" = 0");
414 	return 0;
415 
416 protocol_error:
417 	_leave(" = -EPROTO");
418 	return -EPROTO;
419 }
420 
421 /*
422  * discard hard-ACK'd packets from the Tx window
423  */
rxrpc_rotate_tx_window(struct rxrpc_call * call,u32 hard)424 static void rxrpc_rotate_tx_window(struct rxrpc_call *call, u32 hard)
425 {
426 	unsigned long _skb;
427 	int tail = call->acks_tail, old_tail;
428 	int win = CIRC_CNT(call->acks_head, tail, call->acks_winsz);
429 
430 	_enter("{%u,%u},%u", call->acks_hard, win, hard);
431 
432 	ASSERTCMP(hard - call->acks_hard, <=, win);
433 
434 	while (call->acks_hard < hard) {
435 		smp_read_barrier_depends();
436 		_skb = call->acks_window[tail] & ~1;
437 		rxrpc_free_skb((struct sk_buff *) _skb);
438 		old_tail = tail;
439 		tail = (tail + 1) & (call->acks_winsz - 1);
440 		call->acks_tail = tail;
441 		if (call->acks_unacked == old_tail)
442 			call->acks_unacked = tail;
443 		call->acks_hard++;
444 	}
445 
446 	wake_up(&call->tx_waitq);
447 }
448 
449 /*
450  * clear the Tx window in the event of a failure
451  */
rxrpc_clear_tx_window(struct rxrpc_call * call)452 static void rxrpc_clear_tx_window(struct rxrpc_call *call)
453 {
454 	rxrpc_rotate_tx_window(call, atomic_read(&call->sequence));
455 }
456 
457 /*
458  * drain the out of sequence received packet queue into the packet Rx queue
459  */
rxrpc_drain_rx_oos_queue(struct rxrpc_call * call)460 static int rxrpc_drain_rx_oos_queue(struct rxrpc_call *call)
461 {
462 	struct rxrpc_skb_priv *sp;
463 	struct sk_buff *skb;
464 	bool terminal;
465 	int ret;
466 
467 	_enter("{%d,%d}", call->rx_data_post, call->rx_first_oos);
468 
469 	spin_lock_bh(&call->lock);
470 
471 	ret = -ECONNRESET;
472 	if (test_bit(RXRPC_CALL_RELEASED, &call->flags))
473 		goto socket_unavailable;
474 
475 	skb = skb_dequeue(&call->rx_oos_queue);
476 	if (skb) {
477 		sp = rxrpc_skb(skb);
478 
479 		_debug("drain OOS packet %d [%d]",
480 		       ntohl(sp->hdr.seq), call->rx_first_oos);
481 
482 		if (ntohl(sp->hdr.seq) != call->rx_first_oos) {
483 			skb_queue_head(&call->rx_oos_queue, skb);
484 			call->rx_first_oos = ntohl(rxrpc_skb(skb)->hdr.seq);
485 			_debug("requeue %p {%u}", skb, call->rx_first_oos);
486 		} else {
487 			skb->mark = RXRPC_SKB_MARK_DATA;
488 			terminal = ((sp->hdr.flags & RXRPC_LAST_PACKET) &&
489 				!(sp->hdr.flags & RXRPC_CLIENT_INITIATED));
490 			ret = rxrpc_queue_rcv_skb(call, skb, true, terminal);
491 			BUG_ON(ret < 0);
492 			_debug("drain #%u", call->rx_data_post);
493 			call->rx_data_post++;
494 
495 			/* find out what the next packet is */
496 			skb = skb_peek(&call->rx_oos_queue);
497 			if (skb)
498 				call->rx_first_oos =
499 					ntohl(rxrpc_skb(skb)->hdr.seq);
500 			else
501 				call->rx_first_oos = 0;
502 			_debug("peek %p {%u}", skb, call->rx_first_oos);
503 		}
504 	}
505 
506 	ret = 0;
507 socket_unavailable:
508 	spin_unlock_bh(&call->lock);
509 	_leave(" = %d", ret);
510 	return ret;
511 }
512 
513 /*
514  * insert an out of sequence packet into the buffer
515  */
rxrpc_insert_oos_packet(struct rxrpc_call * call,struct sk_buff * skb)516 static void rxrpc_insert_oos_packet(struct rxrpc_call *call,
517 				    struct sk_buff *skb)
518 {
519 	struct rxrpc_skb_priv *sp, *psp;
520 	struct sk_buff *p;
521 	u32 seq;
522 
523 	sp = rxrpc_skb(skb);
524 	seq = ntohl(sp->hdr.seq);
525 	_enter(",,{%u}", seq);
526 
527 	skb->destructor = rxrpc_packet_destructor;
528 	ASSERTCMP(sp->call, ==, NULL);
529 	sp->call = call;
530 	rxrpc_get_call(call);
531 
532 	/* insert into the buffer in sequence order */
533 	spin_lock_bh(&call->lock);
534 
535 	skb_queue_walk(&call->rx_oos_queue, p) {
536 		psp = rxrpc_skb(p);
537 		if (ntohl(psp->hdr.seq) > seq) {
538 			_debug("insert oos #%u before #%u",
539 			       seq, ntohl(psp->hdr.seq));
540 			skb_insert(p, skb, &call->rx_oos_queue);
541 			goto inserted;
542 		}
543 	}
544 
545 	_debug("append oos #%u", seq);
546 	skb_queue_tail(&call->rx_oos_queue, skb);
547 inserted:
548 
549 	/* we might now have a new front to the queue */
550 	if (call->rx_first_oos == 0 || seq < call->rx_first_oos)
551 		call->rx_first_oos = seq;
552 
553 	read_lock(&call->state_lock);
554 	if (call->state < RXRPC_CALL_COMPLETE &&
555 	    call->rx_data_post == call->rx_first_oos) {
556 		_debug("drain rx oos now");
557 		set_bit(RXRPC_CALL_DRAIN_RX_OOS, &call->events);
558 	}
559 	read_unlock(&call->state_lock);
560 
561 	spin_unlock_bh(&call->lock);
562 	_leave(" [stored #%u]", call->rx_first_oos);
563 }
564 
565 /*
566  * clear the Tx window on final ACK reception
567  */
rxrpc_zap_tx_window(struct rxrpc_call * call)568 static void rxrpc_zap_tx_window(struct rxrpc_call *call)
569 {
570 	struct rxrpc_skb_priv *sp;
571 	struct sk_buff *skb;
572 	unsigned long _skb, *acks_window;
573 	u8 winsz = call->acks_winsz;
574 	int tail;
575 
576 	acks_window = call->acks_window;
577 	call->acks_window = NULL;
578 
579 	while (CIRC_CNT(call->acks_head, call->acks_tail, winsz) > 0) {
580 		tail = call->acks_tail;
581 		smp_read_barrier_depends();
582 		_skb = acks_window[tail] & ~1;
583 		smp_mb();
584 		call->acks_tail = (call->acks_tail + 1) & (winsz - 1);
585 
586 		skb = (struct sk_buff *) _skb;
587 		sp = rxrpc_skb(skb);
588 		_debug("+++ clear Tx %u", ntohl(sp->hdr.seq));
589 		rxrpc_free_skb(skb);
590 	}
591 
592 	kfree(acks_window);
593 }
594 
595 /*
596  * process the extra information that may be appended to an ACK packet
597  */
rxrpc_extract_ackinfo(struct rxrpc_call * call,struct sk_buff * skb,unsigned int latest,int nAcks)598 static void rxrpc_extract_ackinfo(struct rxrpc_call *call, struct sk_buff *skb,
599 				  unsigned int latest, int nAcks)
600 {
601 	struct rxrpc_ackinfo ackinfo;
602 	struct rxrpc_peer *peer;
603 	unsigned int mtu;
604 
605 	if (skb_copy_bits(skb, nAcks + 3, &ackinfo, sizeof(ackinfo)) < 0) {
606 		_leave(" [no ackinfo]");
607 		return;
608 	}
609 
610 	_proto("Rx ACK %%%u Info { rx=%u max=%u rwin=%u jm=%u }",
611 	       latest,
612 	       ntohl(ackinfo.rxMTU), ntohl(ackinfo.maxMTU),
613 	       ntohl(ackinfo.rwind), ntohl(ackinfo.jumbo_max));
614 
615 	mtu = min(ntohl(ackinfo.rxMTU), ntohl(ackinfo.maxMTU));
616 
617 	peer = call->conn->trans->peer;
618 	if (mtu < peer->maxdata) {
619 		spin_lock_bh(&peer->lock);
620 		peer->maxdata = mtu;
621 		peer->mtu = mtu + peer->hdrsize;
622 		spin_unlock_bh(&peer->lock);
623 		_net("Net MTU %u (maxdata %u)", peer->mtu, peer->maxdata);
624 	}
625 }
626 
627 /*
628  * process packets in the reception queue
629  */
rxrpc_process_rx_queue(struct rxrpc_call * call,u32 * _abort_code)630 static int rxrpc_process_rx_queue(struct rxrpc_call *call,
631 				  u32 *_abort_code)
632 {
633 	struct rxrpc_ackpacket ack;
634 	struct rxrpc_skb_priv *sp;
635 	struct sk_buff *skb;
636 	bool post_ACK;
637 	int latest;
638 	u32 hard, tx;
639 
640 	_enter("");
641 
642 process_further:
643 	skb = skb_dequeue(&call->rx_queue);
644 	if (!skb)
645 		return -EAGAIN;
646 
647 	_net("deferred skb %p", skb);
648 
649 	sp = rxrpc_skb(skb);
650 
651 	_debug("process %s [st %d]", rxrpc_pkts[sp->hdr.type], call->state);
652 
653 	post_ACK = false;
654 
655 	switch (sp->hdr.type) {
656 		/* data packets that wind up here have been received out of
657 		 * order, need security processing or are jumbo packets */
658 	case RXRPC_PACKET_TYPE_DATA:
659 		_proto("OOSQ DATA %%%u { #%u }",
660 		       ntohl(sp->hdr.serial), ntohl(sp->hdr.seq));
661 
662 		/* secured packets must be verified and possibly decrypted */
663 		if (rxrpc_verify_packet(call, skb, _abort_code) < 0)
664 			goto protocol_error;
665 
666 		rxrpc_insert_oos_packet(call, skb);
667 		goto process_further;
668 
669 		/* partial ACK to process */
670 	case RXRPC_PACKET_TYPE_ACK:
671 		if (skb_copy_bits(skb, 0, &ack, sizeof(ack)) < 0) {
672 			_debug("extraction failure");
673 			goto protocol_error;
674 		}
675 		if (!skb_pull(skb, sizeof(ack)))
676 			BUG();
677 
678 		latest = ntohl(sp->hdr.serial);
679 		hard = ntohl(ack.firstPacket);
680 		tx = atomic_read(&call->sequence);
681 
682 		_proto("Rx ACK %%%u { m=%hu f=#%u p=#%u s=%%%u r=%s n=%u }",
683 		       latest,
684 		       ntohs(ack.maxSkew),
685 		       hard,
686 		       ntohl(ack.previousPacket),
687 		       ntohl(ack.serial),
688 		       rxrpc_acks(ack.reason),
689 		       ack.nAcks);
690 
691 		rxrpc_extract_ackinfo(call, skb, latest, ack.nAcks);
692 
693 		if (ack.reason == RXRPC_ACK_PING) {
694 			_proto("Rx ACK %%%u PING Request", latest);
695 			rxrpc_propose_ACK(call, RXRPC_ACK_PING_RESPONSE,
696 					  sp->hdr.serial, true);
697 		}
698 
699 		/* discard any out-of-order or duplicate ACKs */
700 		if (latest - call->acks_latest <= 0) {
701 			_debug("discard ACK %d <= %d",
702 			       latest, call->acks_latest);
703 			goto discard;
704 		}
705 		call->acks_latest = latest;
706 
707 		if (call->state != RXRPC_CALL_CLIENT_SEND_REQUEST &&
708 		    call->state != RXRPC_CALL_CLIENT_AWAIT_REPLY &&
709 		    call->state != RXRPC_CALL_SERVER_SEND_REPLY &&
710 		    call->state != RXRPC_CALL_SERVER_AWAIT_ACK)
711 			goto discard;
712 
713 		_debug("Tx=%d H=%u S=%d", tx, call->acks_hard, call->state);
714 
715 		if (hard > 0) {
716 			if (hard - 1 > tx) {
717 				_debug("hard-ACK'd packet %d not transmitted"
718 				       " (%d top)",
719 				       hard - 1, tx);
720 				goto protocol_error;
721 			}
722 
723 			if ((call->state == RXRPC_CALL_CLIENT_AWAIT_REPLY ||
724 			     call->state == RXRPC_CALL_SERVER_AWAIT_ACK) &&
725 			    hard > tx)
726 				goto all_acked;
727 
728 			smp_rmb();
729 			rxrpc_rotate_tx_window(call, hard - 1);
730 		}
731 
732 		if (ack.nAcks > 0) {
733 			if (hard - 1 + ack.nAcks > tx) {
734 				_debug("soft-ACK'd packet %d+%d not"
735 				       " transmitted (%d top)",
736 				       hard - 1, ack.nAcks, tx);
737 				goto protocol_error;
738 			}
739 
740 			if (rxrpc_process_soft_ACKs(call, &ack, skb) < 0)
741 				goto protocol_error;
742 		}
743 		goto discard;
744 
745 		/* complete ACK to process */
746 	case RXRPC_PACKET_TYPE_ACKALL:
747 		goto all_acked;
748 
749 		/* abort and busy are handled elsewhere */
750 	case RXRPC_PACKET_TYPE_BUSY:
751 	case RXRPC_PACKET_TYPE_ABORT:
752 		BUG();
753 
754 		/* connection level events - also handled elsewhere */
755 	case RXRPC_PACKET_TYPE_CHALLENGE:
756 	case RXRPC_PACKET_TYPE_RESPONSE:
757 	case RXRPC_PACKET_TYPE_DEBUG:
758 		BUG();
759 	}
760 
761 	/* if we've had a hard ACK that covers all the packets we've sent, then
762 	 * that ends that phase of the operation */
763 all_acked:
764 	write_lock_bh(&call->state_lock);
765 	_debug("ack all %d", call->state);
766 
767 	switch (call->state) {
768 	case RXRPC_CALL_CLIENT_AWAIT_REPLY:
769 		call->state = RXRPC_CALL_CLIENT_RECV_REPLY;
770 		break;
771 	case RXRPC_CALL_SERVER_AWAIT_ACK:
772 		_debug("srv complete");
773 		call->state = RXRPC_CALL_COMPLETE;
774 		post_ACK = true;
775 		break;
776 	case RXRPC_CALL_CLIENT_SEND_REQUEST:
777 	case RXRPC_CALL_SERVER_RECV_REQUEST:
778 		goto protocol_error_unlock; /* can't occur yet */
779 	default:
780 		write_unlock_bh(&call->state_lock);
781 		goto discard; /* assume packet left over from earlier phase */
782 	}
783 
784 	write_unlock_bh(&call->state_lock);
785 
786 	/* if all the packets we sent are hard-ACK'd, then we can discard
787 	 * whatever we've got left */
788 	_debug("clear Tx %d",
789 	       CIRC_CNT(call->acks_head, call->acks_tail, call->acks_winsz));
790 
791 	del_timer_sync(&call->resend_timer);
792 	clear_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
793 	clear_bit(RXRPC_CALL_RESEND_TIMER, &call->events);
794 
795 	if (call->acks_window)
796 		rxrpc_zap_tx_window(call);
797 
798 	if (post_ACK) {
799 		/* post the final ACK message for userspace to pick up */
800 		_debug("post ACK");
801 		skb->mark = RXRPC_SKB_MARK_FINAL_ACK;
802 		sp->call = call;
803 		rxrpc_get_call(call);
804 		spin_lock_bh(&call->lock);
805 		if (rxrpc_queue_rcv_skb(call, skb, true, true) < 0)
806 			BUG();
807 		spin_unlock_bh(&call->lock);
808 		goto process_further;
809 	}
810 
811 discard:
812 	rxrpc_free_skb(skb);
813 	goto process_further;
814 
815 protocol_error_unlock:
816 	write_unlock_bh(&call->state_lock);
817 protocol_error:
818 	rxrpc_free_skb(skb);
819 	_leave(" = -EPROTO");
820 	return -EPROTO;
821 }
822 
823 /*
824  * post a message to the socket Rx queue for recvmsg() to pick up
825  */
rxrpc_post_message(struct rxrpc_call * call,u32 mark,u32 error,bool fatal)826 static int rxrpc_post_message(struct rxrpc_call *call, u32 mark, u32 error,
827 			      bool fatal)
828 {
829 	struct rxrpc_skb_priv *sp;
830 	struct sk_buff *skb;
831 	int ret;
832 
833 	_enter("{%d,%lx},%u,%u,%d",
834 	       call->debug_id, call->flags, mark, error, fatal);
835 
836 	/* remove timers and things for fatal messages */
837 	if (fatal) {
838 		del_timer_sync(&call->resend_timer);
839 		del_timer_sync(&call->ack_timer);
840 		clear_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
841 	}
842 
843 	if (mark != RXRPC_SKB_MARK_NEW_CALL &&
844 	    !test_bit(RXRPC_CALL_HAS_USERID, &call->flags)) {
845 		_leave("[no userid]");
846 		return 0;
847 	}
848 
849 	if (!test_bit(RXRPC_CALL_TERMINAL_MSG, &call->flags)) {
850 		skb = alloc_skb(0, GFP_NOFS);
851 		if (!skb)
852 			return -ENOMEM;
853 
854 		rxrpc_new_skb(skb);
855 
856 		skb->mark = mark;
857 
858 		sp = rxrpc_skb(skb);
859 		memset(sp, 0, sizeof(*sp));
860 		sp->error = error;
861 		sp->call = call;
862 		rxrpc_get_call(call);
863 
864 		spin_lock_bh(&call->lock);
865 		ret = rxrpc_queue_rcv_skb(call, skb, true, fatal);
866 		spin_unlock_bh(&call->lock);
867 		BUG_ON(ret < 0);
868 	}
869 
870 	return 0;
871 }
872 
873 /*
874  * handle background processing of incoming call packets and ACK / abort
875  * generation
876  */
rxrpc_process_call(struct work_struct * work)877 void rxrpc_process_call(struct work_struct *work)
878 {
879 	struct rxrpc_call *call =
880 		container_of(work, struct rxrpc_call, processor);
881 	struct rxrpc_ackpacket ack;
882 	struct rxrpc_ackinfo ackinfo;
883 	struct rxrpc_header hdr;
884 	struct msghdr msg;
885 	struct kvec iov[5];
886 	unsigned long bits;
887 	__be32 data, pad;
888 	size_t len;
889 	int genbit, loop, nbit, ioc, ret, mtu;
890 	u32 abort_code = RX_PROTOCOL_ERROR;
891 	u8 *acks = NULL;
892 
893 	//printk("\n--------------------\n");
894 	_enter("{%d,%s,%lx} [%lu]",
895 	       call->debug_id, rxrpc_call_states[call->state], call->events,
896 	       (jiffies - call->creation_jif) / (HZ / 10));
897 
898 	if (test_and_set_bit(RXRPC_CALL_PROC_BUSY, &call->flags)) {
899 		_debug("XXXXXXXXXXXXX RUNNING ON MULTIPLE CPUS XXXXXXXXXXXXX");
900 		return;
901 	}
902 
903 	/* there's a good chance we're going to have to send a message, so set
904 	 * one up in advance */
905 	msg.msg_name	= &call->conn->trans->peer->srx.transport.sin;
906 	msg.msg_namelen	= sizeof(call->conn->trans->peer->srx.transport.sin);
907 	msg.msg_control	= NULL;
908 	msg.msg_controllen = 0;
909 	msg.msg_flags	= 0;
910 
911 	hdr.epoch	= call->conn->epoch;
912 	hdr.cid		= call->cid;
913 	hdr.callNumber	= call->call_id;
914 	hdr.seq		= 0;
915 	hdr.type	= RXRPC_PACKET_TYPE_ACK;
916 	hdr.flags	= call->conn->out_clientflag;
917 	hdr.userStatus	= 0;
918 	hdr.securityIndex = call->conn->security_ix;
919 	hdr._rsvd	= 0;
920 	hdr.serviceId	= call->conn->service_id;
921 
922 	memset(iov, 0, sizeof(iov));
923 	iov[0].iov_base	= &hdr;
924 	iov[0].iov_len	= sizeof(hdr);
925 
926 	/* deal with events of a final nature */
927 	if (test_bit(RXRPC_CALL_RELEASE, &call->events)) {
928 		rxrpc_release_call(call);
929 		clear_bit(RXRPC_CALL_RELEASE, &call->events);
930 	}
931 
932 	if (test_bit(RXRPC_CALL_RCVD_ERROR, &call->events)) {
933 		int error;
934 
935 		clear_bit(RXRPC_CALL_CONN_ABORT, &call->events);
936 		clear_bit(RXRPC_CALL_REJECT_BUSY, &call->events);
937 		clear_bit(RXRPC_CALL_ABORT, &call->events);
938 
939 		error = call->conn->trans->peer->net_error;
940 		_debug("post net error %d", error);
941 
942 		if (rxrpc_post_message(call, RXRPC_SKB_MARK_NET_ERROR,
943 				       error, true) < 0)
944 			goto no_mem;
945 		clear_bit(RXRPC_CALL_RCVD_ERROR, &call->events);
946 		goto kill_ACKs;
947 	}
948 
949 	if (test_bit(RXRPC_CALL_CONN_ABORT, &call->events)) {
950 		ASSERTCMP(call->state, >, RXRPC_CALL_COMPLETE);
951 
952 		clear_bit(RXRPC_CALL_REJECT_BUSY, &call->events);
953 		clear_bit(RXRPC_CALL_ABORT, &call->events);
954 
955 		_debug("post conn abort");
956 
957 		if (rxrpc_post_message(call, RXRPC_SKB_MARK_LOCAL_ERROR,
958 				       call->conn->error, true) < 0)
959 			goto no_mem;
960 		clear_bit(RXRPC_CALL_CONN_ABORT, &call->events);
961 		goto kill_ACKs;
962 	}
963 
964 	if (test_bit(RXRPC_CALL_REJECT_BUSY, &call->events)) {
965 		hdr.type = RXRPC_PACKET_TYPE_BUSY;
966 		genbit = RXRPC_CALL_REJECT_BUSY;
967 		goto send_message;
968 	}
969 
970 	if (test_bit(RXRPC_CALL_ABORT, &call->events)) {
971 		ASSERTCMP(call->state, >, RXRPC_CALL_COMPLETE);
972 
973 		if (rxrpc_post_message(call, RXRPC_SKB_MARK_LOCAL_ERROR,
974 				       ECONNABORTED, true) < 0)
975 			goto no_mem;
976 		hdr.type = RXRPC_PACKET_TYPE_ABORT;
977 		data = htonl(call->abort_code);
978 		iov[1].iov_base = &data;
979 		iov[1].iov_len = sizeof(data);
980 		genbit = RXRPC_CALL_ABORT;
981 		goto send_message;
982 	}
983 
984 	if (test_bit(RXRPC_CALL_ACK_FINAL, &call->events)) {
985 		genbit = RXRPC_CALL_ACK_FINAL;
986 
987 		ack.bufferSpace	= htons(8);
988 		ack.maxSkew	= 0;
989 		ack.serial	= 0;
990 		ack.reason	= RXRPC_ACK_IDLE;
991 		ack.nAcks	= 0;
992 		call->ackr_reason = 0;
993 
994 		spin_lock_bh(&call->lock);
995 		ack.serial = call->ackr_serial;
996 		ack.previousPacket = call->ackr_prev_seq;
997 		ack.firstPacket = htonl(call->rx_data_eaten + 1);
998 		spin_unlock_bh(&call->lock);
999 
1000 		pad = 0;
1001 
1002 		iov[1].iov_base = &ack;
1003 		iov[1].iov_len	= sizeof(ack);
1004 		iov[2].iov_base = &pad;
1005 		iov[2].iov_len	= 3;
1006 		iov[3].iov_base = &ackinfo;
1007 		iov[3].iov_len	= sizeof(ackinfo);
1008 		goto send_ACK;
1009 	}
1010 
1011 	if (call->events & ((1 << RXRPC_CALL_RCVD_BUSY) |
1012 			    (1 << RXRPC_CALL_RCVD_ABORT))
1013 	    ) {
1014 		u32 mark;
1015 
1016 		if (test_bit(RXRPC_CALL_RCVD_ABORT, &call->events))
1017 			mark = RXRPC_SKB_MARK_REMOTE_ABORT;
1018 		else
1019 			mark = RXRPC_SKB_MARK_BUSY;
1020 
1021 		_debug("post abort/busy");
1022 		rxrpc_clear_tx_window(call);
1023 		if (rxrpc_post_message(call, mark, ECONNABORTED, true) < 0)
1024 			goto no_mem;
1025 
1026 		clear_bit(RXRPC_CALL_RCVD_BUSY, &call->events);
1027 		clear_bit(RXRPC_CALL_RCVD_ABORT, &call->events);
1028 		goto kill_ACKs;
1029 	}
1030 
1031 	if (test_and_clear_bit(RXRPC_CALL_RCVD_ACKALL, &call->events)) {
1032 		_debug("do implicit ackall");
1033 		rxrpc_clear_tx_window(call);
1034 	}
1035 
1036 	if (test_bit(RXRPC_CALL_LIFE_TIMER, &call->events)) {
1037 		write_lock_bh(&call->state_lock);
1038 		if (call->state <= RXRPC_CALL_COMPLETE) {
1039 			call->state = RXRPC_CALL_LOCALLY_ABORTED;
1040 			call->abort_code = RX_CALL_TIMEOUT;
1041 			set_bit(RXRPC_CALL_ABORT, &call->events);
1042 		}
1043 		write_unlock_bh(&call->state_lock);
1044 
1045 		_debug("post timeout");
1046 		if (rxrpc_post_message(call, RXRPC_SKB_MARK_LOCAL_ERROR,
1047 				       ETIME, true) < 0)
1048 			goto no_mem;
1049 
1050 		clear_bit(RXRPC_CALL_LIFE_TIMER, &call->events);
1051 		goto kill_ACKs;
1052 	}
1053 
1054 	/* deal with assorted inbound messages */
1055 	if (!skb_queue_empty(&call->rx_queue)) {
1056 		switch (rxrpc_process_rx_queue(call, &abort_code)) {
1057 		case 0:
1058 		case -EAGAIN:
1059 			break;
1060 		case -ENOMEM:
1061 			goto no_mem;
1062 		case -EKEYEXPIRED:
1063 		case -EKEYREJECTED:
1064 		case -EPROTO:
1065 			rxrpc_abort_call(call, abort_code);
1066 			goto kill_ACKs;
1067 		}
1068 	}
1069 
1070 	/* handle resending */
1071 	if (test_and_clear_bit(RXRPC_CALL_RESEND_TIMER, &call->events))
1072 		rxrpc_resend_timer(call);
1073 	if (test_and_clear_bit(RXRPC_CALL_RESEND, &call->events))
1074 		rxrpc_resend(call);
1075 
1076 	/* consider sending an ordinary ACK */
1077 	if (test_bit(RXRPC_CALL_ACK, &call->events)) {
1078 		_debug("send ACK: window: %d - %d { %lx }",
1079 		       call->rx_data_eaten, call->ackr_win_top,
1080 		       call->ackr_window[0]);
1081 
1082 		if (call->state > RXRPC_CALL_SERVER_ACK_REQUEST &&
1083 		    call->ackr_reason != RXRPC_ACK_PING_RESPONSE) {
1084 			/* ACK by sending reply DATA packet in this state */
1085 			clear_bit(RXRPC_CALL_ACK, &call->events);
1086 			goto maybe_reschedule;
1087 		}
1088 
1089 		genbit = RXRPC_CALL_ACK;
1090 
1091 		acks = kzalloc(call->ackr_win_top - call->rx_data_eaten,
1092 			       GFP_NOFS);
1093 		if (!acks)
1094 			goto no_mem;
1095 
1096 		//hdr.flags	= RXRPC_SLOW_START_OK;
1097 		ack.bufferSpace	= htons(8);
1098 		ack.maxSkew	= 0;
1099 		ack.serial	= 0;
1100 		ack.reason	= 0;
1101 
1102 		spin_lock_bh(&call->lock);
1103 		ack.reason = call->ackr_reason;
1104 		ack.serial = call->ackr_serial;
1105 		ack.previousPacket = call->ackr_prev_seq;
1106 		ack.firstPacket = htonl(call->rx_data_eaten + 1);
1107 
1108 		ack.nAcks = 0;
1109 		for (loop = 0; loop < RXRPC_ACKR_WINDOW_ASZ; loop++) {
1110 			nbit = loop * BITS_PER_LONG;
1111 			for (bits = call->ackr_window[loop]; bits; bits >>= 1
1112 			     ) {
1113 				_debug("- l=%d n=%d b=%lx", loop, nbit, bits);
1114 				if (bits & 1) {
1115 					acks[nbit] = RXRPC_ACK_TYPE_ACK;
1116 					ack.nAcks = nbit + 1;
1117 				}
1118 				nbit++;
1119 			}
1120 		}
1121 		call->ackr_reason = 0;
1122 		spin_unlock_bh(&call->lock);
1123 
1124 		pad = 0;
1125 
1126 		iov[1].iov_base = &ack;
1127 		iov[1].iov_len	= sizeof(ack);
1128 		iov[2].iov_base = acks;
1129 		iov[2].iov_len	= ack.nAcks;
1130 		iov[3].iov_base = &pad;
1131 		iov[3].iov_len	= 3;
1132 		iov[4].iov_base = &ackinfo;
1133 		iov[4].iov_len	= sizeof(ackinfo);
1134 
1135 		switch (ack.reason) {
1136 		case RXRPC_ACK_REQUESTED:
1137 		case RXRPC_ACK_DUPLICATE:
1138 		case RXRPC_ACK_OUT_OF_SEQUENCE:
1139 		case RXRPC_ACK_EXCEEDS_WINDOW:
1140 		case RXRPC_ACK_NOSPACE:
1141 		case RXRPC_ACK_PING:
1142 		case RXRPC_ACK_PING_RESPONSE:
1143 			goto send_ACK_with_skew;
1144 		case RXRPC_ACK_DELAY:
1145 		case RXRPC_ACK_IDLE:
1146 			goto send_ACK;
1147 		}
1148 	}
1149 
1150 	/* handle completion of security negotiations on an incoming
1151 	 * connection */
1152 	if (test_and_clear_bit(RXRPC_CALL_SECURED, &call->events)) {
1153 		_debug("secured");
1154 		spin_lock_bh(&call->lock);
1155 
1156 		if (call->state == RXRPC_CALL_SERVER_SECURING) {
1157 			_debug("securing");
1158 			write_lock(&call->conn->lock);
1159 			if (!test_bit(RXRPC_CALL_RELEASED, &call->flags) &&
1160 			    !test_bit(RXRPC_CALL_RELEASE, &call->events)) {
1161 				_debug("not released");
1162 				call->state = RXRPC_CALL_SERVER_ACCEPTING;
1163 				list_move_tail(&call->accept_link,
1164 					       &call->socket->acceptq);
1165 			}
1166 			write_unlock(&call->conn->lock);
1167 			read_lock(&call->state_lock);
1168 			if (call->state < RXRPC_CALL_COMPLETE)
1169 				set_bit(RXRPC_CALL_POST_ACCEPT, &call->events);
1170 			read_unlock(&call->state_lock);
1171 		}
1172 
1173 		spin_unlock_bh(&call->lock);
1174 		if (!test_bit(RXRPC_CALL_POST_ACCEPT, &call->events))
1175 			goto maybe_reschedule;
1176 	}
1177 
1178 	/* post a notification of an acceptable connection to the app */
1179 	if (test_bit(RXRPC_CALL_POST_ACCEPT, &call->events)) {
1180 		_debug("post accept");
1181 		if (rxrpc_post_message(call, RXRPC_SKB_MARK_NEW_CALL,
1182 				       0, false) < 0)
1183 			goto no_mem;
1184 		clear_bit(RXRPC_CALL_POST_ACCEPT, &call->events);
1185 		goto maybe_reschedule;
1186 	}
1187 
1188 	/* handle incoming call acceptance */
1189 	if (test_and_clear_bit(RXRPC_CALL_ACCEPTED, &call->events)) {
1190 		_debug("accepted");
1191 		ASSERTCMP(call->rx_data_post, ==, 0);
1192 		call->rx_data_post = 1;
1193 		read_lock_bh(&call->state_lock);
1194 		if (call->state < RXRPC_CALL_COMPLETE)
1195 			set_bit(RXRPC_CALL_DRAIN_RX_OOS, &call->events);
1196 		read_unlock_bh(&call->state_lock);
1197 	}
1198 
1199 	/* drain the out of sequence received packet queue into the packet Rx
1200 	 * queue */
1201 	if (test_and_clear_bit(RXRPC_CALL_DRAIN_RX_OOS, &call->events)) {
1202 		while (call->rx_data_post == call->rx_first_oos)
1203 			if (rxrpc_drain_rx_oos_queue(call) < 0)
1204 				break;
1205 		goto maybe_reschedule;
1206 	}
1207 
1208 	/* other events may have been raised since we started checking */
1209 	goto maybe_reschedule;
1210 
1211 send_ACK_with_skew:
1212 	ack.maxSkew = htons(atomic_read(&call->conn->hi_serial) -
1213 			    ntohl(ack.serial));
1214 send_ACK:
1215 	mtu = call->conn->trans->peer->if_mtu;
1216 	mtu -= call->conn->trans->peer->hdrsize;
1217 	ackinfo.maxMTU	= htonl(mtu);
1218 	ackinfo.rwind	= htonl(rxrpc_rx_window_size);
1219 
1220 	/* permit the peer to send us jumbo packets if it wants to */
1221 	ackinfo.rxMTU	= htonl(rxrpc_rx_mtu);
1222 	ackinfo.jumbo_max = htonl(rxrpc_rx_jumbo_max);
1223 
1224 	hdr.serial = htonl(atomic_inc_return(&call->conn->serial));
1225 	_proto("Tx ACK %%%u { m=%hu f=#%u p=#%u s=%%%u r=%s n=%u }",
1226 	       ntohl(hdr.serial),
1227 	       ntohs(ack.maxSkew),
1228 	       ntohl(ack.firstPacket),
1229 	       ntohl(ack.previousPacket),
1230 	       ntohl(ack.serial),
1231 	       rxrpc_acks(ack.reason),
1232 	       ack.nAcks);
1233 
1234 	del_timer_sync(&call->ack_timer);
1235 	if (ack.nAcks > 0)
1236 		set_bit(RXRPC_CALL_TX_SOFT_ACK, &call->flags);
1237 	goto send_message_2;
1238 
1239 send_message:
1240 	_debug("send message");
1241 
1242 	hdr.serial = htonl(atomic_inc_return(&call->conn->serial));
1243 	_proto("Tx %s %%%u", rxrpc_pkts[hdr.type], ntohl(hdr.serial));
1244 send_message_2:
1245 
1246 	len = iov[0].iov_len;
1247 	ioc = 1;
1248 	if (iov[4].iov_len) {
1249 		ioc = 5;
1250 		len += iov[4].iov_len;
1251 		len += iov[3].iov_len;
1252 		len += iov[2].iov_len;
1253 		len += iov[1].iov_len;
1254 	} else if (iov[3].iov_len) {
1255 		ioc = 4;
1256 		len += iov[3].iov_len;
1257 		len += iov[2].iov_len;
1258 		len += iov[1].iov_len;
1259 	} else if (iov[2].iov_len) {
1260 		ioc = 3;
1261 		len += iov[2].iov_len;
1262 		len += iov[1].iov_len;
1263 	} else if (iov[1].iov_len) {
1264 		ioc = 2;
1265 		len += iov[1].iov_len;
1266 	}
1267 
1268 	ret = kernel_sendmsg(call->conn->trans->local->socket,
1269 			     &msg, iov, ioc, len);
1270 	if (ret < 0) {
1271 		_debug("sendmsg failed: %d", ret);
1272 		read_lock_bh(&call->state_lock);
1273 		if (call->state < RXRPC_CALL_DEAD)
1274 			rxrpc_queue_call(call);
1275 		read_unlock_bh(&call->state_lock);
1276 		goto error;
1277 	}
1278 
1279 	switch (genbit) {
1280 	case RXRPC_CALL_ABORT:
1281 		clear_bit(genbit, &call->events);
1282 		clear_bit(RXRPC_CALL_RCVD_ABORT, &call->events);
1283 		goto kill_ACKs;
1284 
1285 	case RXRPC_CALL_ACK_FINAL:
1286 		write_lock_bh(&call->state_lock);
1287 		if (call->state == RXRPC_CALL_CLIENT_FINAL_ACK)
1288 			call->state = RXRPC_CALL_COMPLETE;
1289 		write_unlock_bh(&call->state_lock);
1290 		goto kill_ACKs;
1291 
1292 	default:
1293 		clear_bit(genbit, &call->events);
1294 		switch (call->state) {
1295 		case RXRPC_CALL_CLIENT_AWAIT_REPLY:
1296 		case RXRPC_CALL_CLIENT_RECV_REPLY:
1297 		case RXRPC_CALL_SERVER_RECV_REQUEST:
1298 		case RXRPC_CALL_SERVER_ACK_REQUEST:
1299 			_debug("start ACK timer");
1300 			rxrpc_propose_ACK(call, RXRPC_ACK_DELAY,
1301 					  call->ackr_serial, false);
1302 		default:
1303 			break;
1304 		}
1305 		goto maybe_reschedule;
1306 	}
1307 
1308 kill_ACKs:
1309 	del_timer_sync(&call->ack_timer);
1310 	if (test_and_clear_bit(RXRPC_CALL_ACK_FINAL, &call->events))
1311 		rxrpc_put_call(call);
1312 	clear_bit(RXRPC_CALL_ACK, &call->events);
1313 
1314 maybe_reschedule:
1315 	if (call->events || !skb_queue_empty(&call->rx_queue)) {
1316 		read_lock_bh(&call->state_lock);
1317 		if (call->state < RXRPC_CALL_DEAD)
1318 			rxrpc_queue_call(call);
1319 		read_unlock_bh(&call->state_lock);
1320 	}
1321 
1322 	/* don't leave aborted connections on the accept queue */
1323 	if (call->state >= RXRPC_CALL_COMPLETE &&
1324 	    !list_empty(&call->accept_link)) {
1325 		_debug("X unlinking once-pending call %p { e=%lx f=%lx c=%x }",
1326 		       call, call->events, call->flags,
1327 		       ntohl(call->conn->cid));
1328 
1329 		read_lock_bh(&call->state_lock);
1330 		if (!test_bit(RXRPC_CALL_RELEASED, &call->flags) &&
1331 		    !test_and_set_bit(RXRPC_CALL_RELEASE, &call->events))
1332 			rxrpc_queue_call(call);
1333 		read_unlock_bh(&call->state_lock);
1334 	}
1335 
1336 error:
1337 	clear_bit(RXRPC_CALL_PROC_BUSY, &call->flags);
1338 	kfree(acks);
1339 
1340 	/* because we don't want two CPUs both processing the work item for one
1341 	 * call at the same time, we use a flag to note when it's busy; however
1342 	 * this means there's a race between clearing the flag and setting the
1343 	 * work pending bit and the work item being processed again */
1344 	if (call->events && !work_pending(&call->processor)) {
1345 		_debug("jumpstart %x", ntohl(call->conn->cid));
1346 		rxrpc_queue_call(call);
1347 	}
1348 
1349 	_leave("");
1350 	return;
1351 
1352 no_mem:
1353 	_debug("out of memory");
1354 	goto maybe_reschedule;
1355 }
1356