?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
inet: Decrease overhead of on-stack inet_cork.
[zen-stable.git] / net / rxrpc / ar-ack.c
blobb6ffe4e1b84afa8a12205bf0bc00ff424341c2c0
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 static unsigned rxrpc_ack_defer = 1;
23
24 static const char *const rxrpc_acks[] = {
25 "---", "REQ", "DUP", "OOS", "WIN", "MEM", "PNG", "PNR", "DLY", "IDL",
26 "-?-"
27 };
28
29 static const s8 rxrpc_ack_priority[] = {
30 [0] = 0,
31 [RXRPC_ACK_DELAY] = 1,
32 [RXRPC_ACK_REQUESTED] = 2,
33 [RXRPC_ACK_IDLE] = 3,
34 [RXRPC_ACK_PING_RESPONSE] = 4,
35 [RXRPC_ACK_DUPLICATE] = 5,
36 [RXRPC_ACK_OUT_OF_SEQUENCE] = 6,
37 [RXRPC_ACK_EXCEEDS_WINDOW] = 7,
38 [RXRPC_ACK_NOSPACE] = 8,
39 };
40
41 /*
42 * propose an ACK be sent
43 */
44 void __rxrpc_propose_ACK(struct rxrpc_call *call, u8 ack_reason,
45 __be32 serial, bool immediate)
46 {
47 unsigned long expiry;
48 s8 prior = rxrpc_ack_priority[ack_reason];
49
50 ASSERTCMP(prior, >, 0);
51
52 _enter("{%d},%s,%%%x,%u",
53 call->debug_id, rxrpc_acks[ack_reason], ntohl(serial),
54 immediate);
55
56 if (prior < rxrpc_ack_priority[call->ackr_reason]) {
57 if (immediate)
58 goto cancel_timer;
59 return;
60 }
61
62 /* update DELAY, IDLE, REQUESTED and PING_RESPONSE ACK serial
63 * numbers */
64 if (prior == rxrpc_ack_priority[call->ackr_reason]) {
65 if (prior <= 4)
66 call->ackr_serial = serial;
67 if (immediate)
68 goto cancel_timer;
69 return;
70 }
71
72 call->ackr_reason = ack_reason;
73 call->ackr_serial = serial;
74
75 switch (ack_reason) {
76 case RXRPC_ACK_DELAY:
77 _debug("run delay timer");
78 call->ack_timer.expires = jiffies + rxrpc_ack_timeout * HZ;
79 add_timer(&call->ack_timer);
80 return;
81
82 case RXRPC_ACK_IDLE:
83 if (!immediate) {
84 _debug("run defer timer");
85 expiry = 1;
86 goto run_timer;
87 }
88 goto cancel_timer;
89
90 case RXRPC_ACK_REQUESTED:
91 if (!rxrpc_ack_defer)
92 goto cancel_timer;
93 if (!immediate || serial == cpu_to_be32(1)) {
94 _debug("run defer timer");
95 expiry = rxrpc_ack_defer;
96 goto run_timer;
97 }
98
99 default:
100 _debug("immediate ACK");
101 goto cancel_timer;
102 }
103
104 run_timer:
105 expiry += jiffies;
106 if (!timer_pending(&call->ack_timer) ||
107 time_after(call->ack_timer.expires, expiry))
108 mod_timer(&call->ack_timer, expiry);
109 return;
110
111 cancel_timer:
112 _debug("cancel timer %%%u", ntohl(serial));
113 try_to_del_timer_sync(&call->ack_timer);
114 read_lock_bh(&call->state_lock);
115 if (call->state <= RXRPC_CALL_COMPLETE &&
116 !test_and_set_bit(RXRPC_CALL_ACK, &call->events))
117 rxrpc_queue_call(call);
118 read_unlock_bh(&call->state_lock);
119 }
120
121 /*
122 * propose an ACK be sent, locking the call structure
123 */
124 void rxrpc_propose_ACK(struct rxrpc_call *call, u8 ack_reason,
125 __be32 serial, bool immediate)
126 {
127 s8 prior = rxrpc_ack_priority[ack_reason];
128
129 if (prior > rxrpc_ack_priority[call->ackr_reason]) {
130 spin_lock_bh(&call->lock);
131 __rxrpc_propose_ACK(call, ack_reason, serial, immediate);
132 spin_unlock_bh(&call->lock);
133 }
134 }
135
136 /*
137 * set the resend timer
138 */
139 static void rxrpc_set_resend(struct rxrpc_call *call, u8 resend,
140 unsigned long resend_at)
141 {
142 read_lock_bh(&call->state_lock);
143 if (call->state >= RXRPC_CALL_COMPLETE)
144 resend = 0;
145
146 if (resend & 1) {
147 _debug("SET RESEND");
148 set_bit(RXRPC_CALL_RESEND, &call->events);
149 }
150
151 if (resend & 2) {
152 _debug("MODIFY RESEND TIMER");
153 set_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
154 mod_timer(&call->resend_timer, resend_at);
155 } else {
156 _debug("KILL RESEND TIMER");
157 del_timer_sync(&call->resend_timer);
158 clear_bit(RXRPC_CALL_RESEND_TIMER, &call->events);
159 clear_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
160 }
161 read_unlock_bh(&call->state_lock);
162 }
163
164 /*
165 * resend packets
166 */
167 static void rxrpc_resend(struct rxrpc_call *call)
168 {
169 struct rxrpc_skb_priv *sp;
170 struct rxrpc_header *hdr;
171 struct sk_buff *txb;
172 unsigned long *p_txb, resend_at;
173 int loop, stop;
174 u8 resend;
175
176 _enter("{%d,%d,%d,%d},",
177 call->acks_hard, call->acks_unacked,
178 atomic_read(&call->sequence),
179 CIRC_CNT(call->acks_head, call->acks_tail, call->acks_winsz));
180
181 stop = 0;
182 resend = 0;
183 resend_at = 0;
184
185 for (loop = call->acks_tail;
186 loop != call->acks_head || stop;
187 loop = (loop + 1) & (call->acks_winsz - 1)
188 ) {
189 p_txb = call->acks_window + loop;
190 smp_read_barrier_depends();
191 if (*p_txb & 1)
192 continue;
193
194 txb = (struct sk_buff *) *p_txb;
195 sp = rxrpc_skb(txb);
196
197 if (sp->need_resend) {
198 sp->need_resend = 0;
199
200 /* each Tx packet has a new serial number */
201 sp->hdr.serial =
202 htonl(atomic_inc_return(&call->conn->serial));
203
204 hdr = (struct rxrpc_header *) txb->head;
205 hdr->serial = sp->hdr.serial;
206
207 _proto("Tx DATA %%%u { #%d }",
208 ntohl(sp->hdr.serial), ntohl(sp->hdr.seq));
209 if (rxrpc_send_packet(call->conn->trans, txb) < 0) {
210 stop = 0;
211 sp->resend_at = jiffies + 3;
212 } else {
213 sp->resend_at =
214 jiffies + rxrpc_resend_timeout * HZ;
215 }
216 }
217
218 if (time_after_eq(jiffies + 1, sp->resend_at)) {
219 sp->need_resend = 1;
220 resend |= 1;
221 } else if (resend & 2) {
222 if (time_before(sp->resend_at, resend_at))
223 resend_at = sp->resend_at;
224 } else {
225 resend_at = sp->resend_at;
226 resend |= 2;
227 }
228 }
229
230 rxrpc_set_resend(call, resend, resend_at);
231 _leave("");
232 }
233
234 /*
235 * handle resend timer expiry
236 */
237 static void rxrpc_resend_timer(struct rxrpc_call *call)
238 {
239 struct rxrpc_skb_priv *sp;
240 struct sk_buff *txb;
241 unsigned long *p_txb, resend_at;
242 int loop;
243 u8 resend;
244
245 _enter("%d,%d,%d",
246 call->acks_tail, call->acks_unacked, call->acks_head);
247
248 if (call->state >= RXRPC_CALL_COMPLETE)
249 return;
250
251 resend = 0;
252 resend_at = 0;
253
254 for (loop = call->acks_unacked;
255 loop != call->acks_head;
256 loop = (loop + 1) & (call->acks_winsz - 1)
257 ) {
258 p_txb = call->acks_window + loop;
259 smp_read_barrier_depends();
260 txb = (struct sk_buff *) (*p_txb & ~1);
261 sp = rxrpc_skb(txb);
262
263 ASSERT(!(*p_txb & 1));
264
265 if (sp->need_resend) {
266 ;
267 } else if (time_after_eq(jiffies + 1, sp->resend_at)) {
268 sp->need_resend = 1;
269 resend |= 1;
270 } else if (resend & 2) {
271 if (time_before(sp->resend_at, resend_at))
272 resend_at = sp->resend_at;
273 } else {
274 resend_at = sp->resend_at;
275 resend |= 2;
276 }
277 }
278
279 rxrpc_set_resend(call, resend, resend_at);
280 _leave("");
281 }
282
283 /*
284 * process soft ACKs of our transmitted packets
285 * - these indicate packets the peer has or has not received, but hasn't yet
286 * given to the consumer, and so can still be discarded and re-requested
287 */
288 static int rxrpc_process_soft_ACKs(struct rxrpc_call *call,
289 struct rxrpc_ackpacket *ack,
290 struct sk_buff *skb)
291 {
292 struct rxrpc_skb_priv *sp;
293 struct sk_buff *txb;
294 unsigned long *p_txb, resend_at;
295 int loop;
296 u8 sacks[RXRPC_MAXACKS], resend;
297
298 _enter("{%d,%d},{%d},",
299 call->acks_hard,
300 CIRC_CNT(call->acks_head, call->acks_tail, call->acks_winsz),
301 ack->nAcks);
302
303 if (skb_copy_bits(skb, 0, sacks, ack->nAcks) < 0)
304 goto protocol_error;
305
306 resend = 0;
307 resend_at = 0;
308 for (loop = 0; loop < ack->nAcks; loop++) {
309 p_txb = call->acks_window;
310 p_txb += (call->acks_tail + loop) & (call->acks_winsz - 1);
311 smp_read_barrier_depends();
312 txb = (struct sk_buff *) (*p_txb & ~1);
313 sp = rxrpc_skb(txb);
314
315 switch (sacks[loop]) {
316 case RXRPC_ACK_TYPE_ACK:
317 sp->need_resend = 0;
318 *p_txb |= 1;
319 break;
320 case RXRPC_ACK_TYPE_NACK:
321 sp->need_resend = 1;
322 *p_txb &= ~1;
323 resend = 1;
324 break;
325 default:
326 _debug("Unsupported ACK type %d", sacks[loop]);
327 goto protocol_error;
328 }
329 }
330
331 smp_mb();
332 call->acks_unacked = (call->acks_tail + loop) & (call->acks_winsz - 1);
333
334 /* anything not explicitly ACK'd is implicitly NACK'd, but may just not
335 * have been received or processed yet by the far end */
336 for (loop = call->acks_unacked;
337 loop != call->acks_head;
338 loop = (loop + 1) & (call->acks_winsz - 1)
339 ) {
340 p_txb = call->acks_window + loop;
341 smp_read_barrier_depends();
342 txb = (struct sk_buff *) (*p_txb & ~1);
343 sp = rxrpc_skb(txb);
344
345 if (*p_txb & 1) {
346 /* packet must have been discarded */
347 sp->need_resend = 1;
348 *p_txb &= ~1;
349 resend |= 1;
350 } else if (sp->need_resend) {
351 ;
352 } else if (time_after_eq(jiffies + 1, sp->resend_at)) {
353 sp->need_resend = 1;
354 resend |= 1;
355 } else if (resend & 2) {
356 if (time_before(sp->resend_at, resend_at))
357 resend_at = sp->resend_at;
358 } else {
359 resend_at = sp->resend_at;
360 resend |= 2;
361 }
362 }
363
364 rxrpc_set_resend(call, resend, resend_at);
365 _leave(" = 0");
366 return 0;
367
368 protocol_error:
369 _leave(" = -EPROTO");
370 return -EPROTO;
371 }
372
373 /*
374 * discard hard-ACK'd packets from the Tx window
375 */
376 static void rxrpc_rotate_tx_window(struct rxrpc_call *call, u32 hard)
377 {
378 struct rxrpc_skb_priv *sp;
379 unsigned long _skb;
380 int tail = call->acks_tail, old_tail;
381 int win = CIRC_CNT(call->acks_head, tail, call->acks_winsz);
382
383 _enter("{%u,%u},%u", call->acks_hard, win, hard);
384
385 ASSERTCMP(hard - call->acks_hard, <=, win);
386
387 while (call->acks_hard < hard) {
388 smp_read_barrier_depends();
389 _skb = call->acks_window[tail] & ~1;
390 sp = rxrpc_skb((struct sk_buff *) _skb);
391 rxrpc_free_skb((struct sk_buff *) _skb);
392 old_tail = tail;
393 tail = (tail + 1) & (call->acks_winsz - 1);
394 call->acks_tail = tail;
395 if (call->acks_unacked == old_tail)
396 call->acks_unacked = tail;
397 call->acks_hard++;
398 }
399
400 wake_up(&call->tx_waitq);
401 }
402
403 /*
404 * clear the Tx window in the event of a failure
405 */
406 static void rxrpc_clear_tx_window(struct rxrpc_call *call)
407 {
408 rxrpc_rotate_tx_window(call, atomic_read(&call->sequence));
409 }
410
411 /*
412 * drain the out of sequence received packet queue into the packet Rx queue
413 */
414 static int rxrpc_drain_rx_oos_queue(struct rxrpc_call *call)
415 {
416 struct rxrpc_skb_priv *sp;
417 struct sk_buff *skb;
418 bool terminal;
419 int ret;
420
421 _enter("{%d,%d}", call->rx_data_post, call->rx_first_oos);
422
423 spin_lock_bh(&call->lock);
424
425 ret = -ECONNRESET;
426 if (test_bit(RXRPC_CALL_RELEASED, &call->flags))
427 goto socket_unavailable;
428
429 skb = skb_dequeue(&call->rx_oos_queue);
430 if (skb) {
431 sp = rxrpc_skb(skb);
432
433 _debug("drain OOS packet %d [%d]",
434 ntohl(sp->hdr.seq), call->rx_first_oos);
435
436 if (ntohl(sp->hdr.seq) != call->rx_first_oos) {
437 skb_queue_head(&call->rx_oos_queue, skb);
438 call->rx_first_oos = ntohl(rxrpc_skb(skb)->hdr.seq);
439 _debug("requeue %p {%u}", skb, call->rx_first_oos);
440 } else {
441 skb->mark = RXRPC_SKB_MARK_DATA;
442 terminal = ((sp->hdr.flags & RXRPC_LAST_PACKET) &&
443 !(sp->hdr.flags & RXRPC_CLIENT_INITIATED));
444 ret = rxrpc_queue_rcv_skb(call, skb, true, terminal);
445 BUG_ON(ret < 0);
446 _debug("drain #%u", call->rx_data_post);
447 call->rx_data_post++;
448
449 /* find out what the next packet is */
450 skb = skb_peek(&call->rx_oos_queue);
451 if (skb)
452 call->rx_first_oos =
453 ntohl(rxrpc_skb(skb)->hdr.seq);
454 else
455 call->rx_first_oos = 0;
456 _debug("peek %p {%u}", skb, call->rx_first_oos);
457 }
458 }
459
460 ret = 0;
461 socket_unavailable:
462 spin_unlock_bh(&call->lock);
463 _leave(" = %d", ret);
464 return ret;
465 }
466
467 /*
468 * insert an out of sequence packet into the buffer
469 */
470 static void rxrpc_insert_oos_packet(struct rxrpc_call *call,
471 struct sk_buff *skb)
472 {
473 struct rxrpc_skb_priv *sp, *psp;
474 struct sk_buff *p;
475 u32 seq;
476
477 sp = rxrpc_skb(skb);
478 seq = ntohl(sp->hdr.seq);
479 _enter(",,{%u}", seq);
480
481 skb->destructor = rxrpc_packet_destructor;
482 ASSERTCMP(sp->call, ==, NULL);
483 sp->call = call;
484 rxrpc_get_call(call);
485
486 /* insert into the buffer in sequence order */
487 spin_lock_bh(&call->lock);
488
489 skb_queue_walk(&call->rx_oos_queue, p) {
490 psp = rxrpc_skb(p);
491 if (ntohl(psp->hdr.seq) > seq) {
492 _debug("insert oos #%u before #%u",
493 seq, ntohl(psp->hdr.seq));
494 skb_insert(p, skb, &call->rx_oos_queue);
495 goto inserted;
496 }
497 }
498
499 _debug("append oos #%u", seq);
500 skb_queue_tail(&call->rx_oos_queue, skb);
501 inserted:
502
503 /* we might now have a new front to the queue */
504 if (call->rx_first_oos == 0 || seq < call->rx_first_oos)
505 call->rx_first_oos = seq;
506
507 read_lock(&call->state_lock);
508 if (call->state < RXRPC_CALL_COMPLETE &&
509 call->rx_data_post == call->rx_first_oos) {
510 _debug("drain rx oos now");
511 set_bit(RXRPC_CALL_DRAIN_RX_OOS, &call->events);
512 }
513 read_unlock(&call->state_lock);
514
515 spin_unlock_bh(&call->lock);
516 _leave(" [stored #%u]", call->rx_first_oos);
517 }
518
519 /*
520 * clear the Tx window on final ACK reception
521 */
522 static void rxrpc_zap_tx_window(struct rxrpc_call *call)
523 {
524 struct rxrpc_skb_priv *sp;
525 struct sk_buff *skb;
526 unsigned long _skb, *acks_window;
527 u8 winsz = call->acks_winsz;
528 int tail;
529
530 acks_window = call->acks_window;
531 call->acks_window = NULL;
532
533 while (CIRC_CNT(call->acks_head, call->acks_tail, winsz) > 0) {
534 tail = call->acks_tail;
535 smp_read_barrier_depends();
536 _skb = acks_window[tail] & ~1;
537 smp_mb();
538 call->acks_tail = (call->acks_tail + 1) & (winsz - 1);
539
540 skb = (struct sk_buff *) _skb;
541 sp = rxrpc_skb(skb);
542 _debug("+++ clear Tx %u", ntohl(sp->hdr.seq));
543 rxrpc_free_skb(skb);
544 }
545
546 kfree(acks_window);
547 }
548
549 /*
550 * process the extra information that may be appended to an ACK packet
551 */
552 static void rxrpc_extract_ackinfo(struct rxrpc_call *call, struct sk_buff *skb,
553 unsigned latest, int nAcks)
554 {
555 struct rxrpc_ackinfo ackinfo;
556 struct rxrpc_peer *peer;
557 unsigned mtu;
558
559 if (skb_copy_bits(skb, nAcks + 3, &ackinfo, sizeof(ackinfo)) < 0) {
560 _leave(" [no ackinfo]");
561 return;
562 }
563
564 _proto("Rx ACK %%%u Info { rx=%u max=%u rwin=%u jm=%u }",
565 latest,
566 ntohl(ackinfo.rxMTU), ntohl(ackinfo.maxMTU),
567 ntohl(ackinfo.rwind), ntohl(ackinfo.jumbo_max));
568
569 mtu = min(ntohl(ackinfo.rxMTU), ntohl(ackinfo.maxMTU));
570
571 peer = call->conn->trans->peer;
572 if (mtu < peer->maxdata) {
573 spin_lock_bh(&peer->lock);
574 peer->maxdata = mtu;
575 peer->mtu = mtu + peer->hdrsize;
576 spin_unlock_bh(&peer->lock);
577 _net("Net MTU %u (maxdata %u)", peer->mtu, peer->maxdata);
578 }
579 }
580
581 /*
582 * process packets in the reception queue
583 */
584 static int rxrpc_process_rx_queue(struct rxrpc_call *call,
585 u32 *_abort_code)
586 {
587 struct rxrpc_ackpacket ack;
588 struct rxrpc_skb_priv *sp;
589 struct sk_buff *skb;
590 bool post_ACK;
591 int latest;
592 u32 hard, tx;
593
594 _enter("");
595
596 process_further:
597 skb = skb_dequeue(&call->rx_queue);
598 if (!skb)
599 return -EAGAIN;
600
601 _net("deferred skb %p", skb);
602
603 sp = rxrpc_skb(skb);
604
605 _debug("process %s [st %d]", rxrpc_pkts[sp->hdr.type], call->state);
606
607 post_ACK = false;
608
609 switch (sp->hdr.type) {
610 /* data packets that wind up here have been received out of
611 * order, need security processing or are jumbo packets */
612 case RXRPC_PACKET_TYPE_DATA:
613 _proto("OOSQ DATA %%%u { #%u }",
614 ntohl(sp->hdr.serial), ntohl(sp->hdr.seq));
615
616 /* secured packets must be verified and possibly decrypted */
617 if (rxrpc_verify_packet(call, skb, _abort_code) < 0)
618 goto protocol_error;
619
620 rxrpc_insert_oos_packet(call, skb);
621 goto process_further;
622
623 /* partial ACK to process */
624 case RXRPC_PACKET_TYPE_ACK:
625 if (skb_copy_bits(skb, 0, &ack, sizeof(ack)) < 0) {
626 _debug("extraction failure");
627 goto protocol_error;
628 }
629 if (!skb_pull(skb, sizeof(ack)))
630 BUG();
631
632 latest = ntohl(sp->hdr.serial);
633 hard = ntohl(ack.firstPacket);
634 tx = atomic_read(&call->sequence);
635
636 _proto("Rx ACK %%%u { m=%hu f=#%u p=#%u s=%%%u r=%s n=%u }",
637 latest,
638 ntohs(ack.maxSkew),
639 hard,
640 ntohl(ack.previousPacket),
641 ntohl(ack.serial),
642 rxrpc_acks[ack.reason],
643 ack.nAcks);
644
645 rxrpc_extract_ackinfo(call, skb, latest, ack.nAcks);
646
647 if (ack.reason == RXRPC_ACK_PING) {
648 _proto("Rx ACK %%%u PING Request", latest);
649 rxrpc_propose_ACK(call, RXRPC_ACK_PING_RESPONSE,
650 sp->hdr.serial, true);
651 }
652
653 /* discard any out-of-order or duplicate ACKs */
654 if (latest - call->acks_latest <= 0) {
655 _debug("discard ACK %d <= %d",
656 latest, call->acks_latest);
657 goto discard;
658 }
659 call->acks_latest = latest;
660
661 if (call->state != RXRPC_CALL_CLIENT_SEND_REQUEST &&
662 call->state != RXRPC_CALL_CLIENT_AWAIT_REPLY &&
663 call->state != RXRPC_CALL_SERVER_SEND_REPLY &&
664 call->state != RXRPC_CALL_SERVER_AWAIT_ACK)
665 goto discard;
666
667 _debug("Tx=%d H=%u S=%d", tx, call->acks_hard, call->state);
668
669 if (hard > 0) {
670 if (hard - 1 > tx) {
671 _debug("hard-ACK'd packet %d not transmitted"
672 " (%d top)",
673 hard - 1, tx);
674 goto protocol_error;
675 }
676
677 if ((call->state == RXRPC_CALL_CLIENT_AWAIT_REPLY ||
678 call->state == RXRPC_CALL_SERVER_AWAIT_ACK) &&
679 hard > tx)
680 goto all_acked;
681
682 smp_rmb();
683 rxrpc_rotate_tx_window(call, hard - 1);
684 }
685
686 if (ack.nAcks > 0) {
687 if (hard - 1 + ack.nAcks > tx) {
688 _debug("soft-ACK'd packet %d+%d not"
689 " transmitted (%d top)",
690 hard - 1, ack.nAcks, tx);
691 goto protocol_error;
692 }
693
694 if (rxrpc_process_soft_ACKs(call, &ack, skb) < 0)
695 goto protocol_error;
696 }
697 goto discard;
698
699 /* complete ACK to process */
700 case RXRPC_PACKET_TYPE_ACKALL:
701 goto all_acked;
702
703 /* abort and busy are handled elsewhere */
704 case RXRPC_PACKET_TYPE_BUSY:
705 case RXRPC_PACKET_TYPE_ABORT:
706 BUG();
707
708 /* connection level events - also handled elsewhere */
709 case RXRPC_PACKET_TYPE_CHALLENGE:
710 case RXRPC_PACKET_TYPE_RESPONSE:
711 case RXRPC_PACKET_TYPE_DEBUG:
712 BUG();
713 }
714
715 /* if we've had a hard ACK that covers all the packets we've sent, then
716 * that ends that phase of the operation */
717 all_acked:
718 write_lock_bh(&call->state_lock);
719 _debug("ack all %d", call->state);
720
721 switch (call->state) {
722 case RXRPC_CALL_CLIENT_AWAIT_REPLY:
723 call->state = RXRPC_CALL_CLIENT_RECV_REPLY;
724 break;
725 case RXRPC_CALL_SERVER_AWAIT_ACK:
726 _debug("srv complete");
727 call->state = RXRPC_CALL_COMPLETE;
728 post_ACK = true;
729 break;
730 case RXRPC_CALL_CLIENT_SEND_REQUEST:
731 case RXRPC_CALL_SERVER_RECV_REQUEST:
732 goto protocol_error_unlock; /* can't occur yet */
733 default:
734 write_unlock_bh(&call->state_lock);
735 goto discard; /* assume packet left over from earlier phase */
736 }
737
738 write_unlock_bh(&call->state_lock);
739
740 /* if all the packets we sent are hard-ACK'd, then we can discard
741 * whatever we've got left */
742 _debug("clear Tx %d",
743 CIRC_CNT(call->acks_head, call->acks_tail, call->acks_winsz));
744
745 del_timer_sync(&call->resend_timer);
746 clear_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
747 clear_bit(RXRPC_CALL_RESEND_TIMER, &call->events);
748
749 if (call->acks_window)
750 rxrpc_zap_tx_window(call);
751
752 if (post_ACK) {
753 /* post the final ACK message for userspace to pick up */
754 _debug("post ACK");
755 skb->mark = RXRPC_SKB_MARK_FINAL_ACK;
756 sp->call = call;
757 rxrpc_get_call(call);
758 spin_lock_bh(&call->lock);
759 if (rxrpc_queue_rcv_skb(call, skb, true, true) < 0)
760 BUG();
761 spin_unlock_bh(&call->lock);
762 goto process_further;
763 }
764
765 discard:
766 rxrpc_free_skb(skb);
767 goto process_further;
768
769 protocol_error_unlock:
770 write_unlock_bh(&call->state_lock);
771 protocol_error:
772 rxrpc_free_skb(skb);
773 _leave(" = -EPROTO");
774 return -EPROTO;
775 }
776
777 /*
778 * post a message to the socket Rx queue for recvmsg() to pick up
779 */
780 static int rxrpc_post_message(struct rxrpc_call *call, u32 mark, u32 error,
781 bool fatal)
782 {
783 struct rxrpc_skb_priv *sp;
784 struct sk_buff *skb;
785 int ret;
786
787 _enter("{%d,%lx},%u,%u,%d",
788 call->debug_id, call->flags, mark, error, fatal);
789
790 /* remove timers and things for fatal messages */
791 if (fatal) {
792 del_timer_sync(&call->resend_timer);
793 del_timer_sync(&call->ack_timer);
794 clear_bit(RXRPC_CALL_RUN_RTIMER, &call->flags);
795 }
796
797 if (mark != RXRPC_SKB_MARK_NEW_CALL &&
798 !test_bit(RXRPC_CALL_HAS_USERID, &call->flags)) {
799 _leave("[no userid]");
800 return 0;
801 }
802
803 if (!test_bit(RXRPC_CALL_TERMINAL_MSG, &call->flags)) {
804 skb = alloc_skb(0, GFP_NOFS);
805 if (!skb)
806 return -ENOMEM;
807
808 rxrpc_new_skb(skb);
809
810 skb->mark = mark;
811
812 sp = rxrpc_skb(skb);
813 memset(sp, 0, sizeof(*sp));
814 sp->error = error;
815 sp->call = call;
816 rxrpc_get_call(call);
817
818 spin_lock_bh(&call->lock);
819 ret = rxrpc_queue_rcv_skb(call, skb, true, fatal);
820 spin_unlock_bh(&call->lock);
821 BUG_ON(ret < 0);
822 }
823
824 return 0;
825 }
826
827 /*
828 * handle background processing of incoming call packets and ACK / abort
829 * generation
830 */
831 void rxrpc_process_call(struct work_struct *work)
832 {
833 struct rxrpc_call *call =
834 container_of(work, struct rxrpc_call, processor);
835 struct rxrpc_ackpacket ack;
836 struct rxrpc_ackinfo ackinfo;
837 struct rxrpc_header hdr;
838 struct msghdr msg;
839 struct kvec iov[5];
840 unsigned long bits;
841 __be32 data, pad;
842 size_t len;
843 int genbit, loop, nbit, ioc, ret, mtu;
844 u32 abort_code = RX_PROTOCOL_ERROR;
845 u8 *acks = NULL;
846
847 //printk("\n--------------------\n");
848 _enter("{%d,%s,%lx} [%lu]",
849 call->debug_id, rxrpc_call_states[call->state], call->events,
850 (jiffies - call->creation_jif) / (HZ / 10));
851
852 if (test_and_set_bit(RXRPC_CALL_PROC_BUSY, &call->flags)) {
853 _debug("XXXXXXXXXXXXX RUNNING ON MULTIPLE CPUS XXXXXXXXXXXXX");
854 return;
855 }
856
857 /* there's a good chance we're going to have to send a message, so set
858 * one up in advance */
859 msg.msg_name = &call->conn->trans->peer->srx.transport.sin;
860 msg.msg_namelen = sizeof(call->conn->trans->peer->srx.transport.sin);
861 msg.msg_control = NULL;
862 msg.msg_controllen = 0;
863 msg.msg_flags = 0;
864
865 hdr.epoch = call->conn->epoch;
866 hdr.cid = call->cid;
867 hdr.callNumber = call->call_id;
868 hdr.seq = 0;
869 hdr.type = RXRPC_PACKET_TYPE_ACK;
870 hdr.flags = call->conn->out_clientflag;
871 hdr.userStatus = 0;
872 hdr.securityIndex = call->conn->security_ix;
873 hdr._rsvd = 0;
874 hdr.serviceId = call->conn->service_id;
875
876 memset(iov, 0, sizeof(iov));
877 iov[0].iov_base = &hdr;
878 iov[0].iov_len = sizeof(hdr);
879
880 /* deal with events of a final nature */
881 if (test_bit(RXRPC_CALL_RELEASE, &call->events)) {
882 rxrpc_release_call(call);
883 clear_bit(RXRPC_CALL_RELEASE, &call->events);
884 }
885
886 if (test_bit(RXRPC_CALL_RCVD_ERROR, &call->events)) {
887 int error;
888
889 clear_bit(RXRPC_CALL_CONN_ABORT, &call->events);
890 clear_bit(RXRPC_CALL_REJECT_BUSY, &call->events);
891 clear_bit(RXRPC_CALL_ABORT, &call->events);
892
893 error = call->conn->trans->peer->net_error;
894 _debug("post net error %d", error);
895
896 if (rxrpc_post_message(call, RXRPC_SKB_MARK_NET_ERROR,
897 error, true) < 0)
898 goto no_mem;
899 clear_bit(RXRPC_CALL_RCVD_ERROR, &call->events);
900 goto kill_ACKs;
901 }
902
903 if (test_bit(RXRPC_CALL_CONN_ABORT, &call->events)) {
904 ASSERTCMP(call->state, >, RXRPC_CALL_COMPLETE);
905
906 clear_bit(RXRPC_CALL_REJECT_BUSY, &call->events);
907 clear_bit(RXRPC_CALL_ABORT, &call->events);
908
909 _debug("post conn abort");
910
911 if (rxrpc_post_message(call, RXRPC_SKB_MARK_LOCAL_ERROR,
912 call->conn->error, true) < 0)
913 goto no_mem;
914 clear_bit(RXRPC_CALL_CONN_ABORT, &call->events);
915 goto kill_ACKs;
916 }
917
918 if (test_bit(RXRPC_CALL_REJECT_BUSY, &call->events)) {
919 hdr.type = RXRPC_PACKET_TYPE_BUSY;
920 genbit = RXRPC_CALL_REJECT_BUSY;
921 goto send_message;
922 }
923
924 if (test_bit(RXRPC_CALL_ABORT, &call->events)) {
925 ASSERTCMP(call->state, >, RXRPC_CALL_COMPLETE);
926
927 if (rxrpc_post_message(call, RXRPC_SKB_MARK_LOCAL_ERROR,
928 ECONNABORTED, true) < 0)
929 goto no_mem;
930 hdr.type = RXRPC_PACKET_TYPE_ABORT;
931 data = htonl(call->abort_code);
932 iov[1].iov_base = &data;
933 iov[1].iov_len = sizeof(data);
934 genbit = RXRPC_CALL_ABORT;
935 goto send_message;
936 }
937
938 if (test_bit(RXRPC_CALL_ACK_FINAL, &call->events)) {
939 genbit = RXRPC_CALL_ACK_FINAL;
940
941 ack.bufferSpace = htons(8);
942 ack.maxSkew = 0;
943 ack.serial = 0;
944 ack.reason = RXRPC_ACK_IDLE;
945 ack.nAcks = 0;
946 call->ackr_reason = 0;
947
948 spin_lock_bh(&call->lock);
949 ack.serial = call->ackr_serial;
950 ack.previousPacket = call->ackr_prev_seq;
951 ack.firstPacket = htonl(call->rx_data_eaten + 1);
952 spin_unlock_bh(&call->lock);
953
954 pad = 0;
955
956 iov[1].iov_base = &ack;
957 iov[1].iov_len = sizeof(ack);
958 iov[2].iov_base = &pad;
959 iov[2].iov_len = 3;
960 iov[3].iov_base = &ackinfo;
961 iov[3].iov_len = sizeof(ackinfo);
962 goto send_ACK;
963 }
964
965 if (call->events & ((1 << RXRPC_CALL_RCVD_BUSY) |
966 (1 << RXRPC_CALL_RCVD_ABORT))
967 ) {
968 u32 mark;
969
970 if (test_bit(RXRPC_CALL_RCVD_ABORT, &call->events))
971 mark = RXRPC_SKB_MARK_REMOTE_ABORT;
972 else
973 mark = RXRPC_SKB_MARK_BUSY;
974
975 _debug("post abort/busy");
976 rxrpc_clear_tx_window(call);
977 if (rxrpc_post_message(call, mark, ECONNABORTED, true) < 0)
978 goto no_mem;
979
980 clear_bit(RXRPC_CALL_RCVD_BUSY, &call->events);
981 clear_bit(RXRPC_CALL_RCVD_ABORT, &call->events);
982 goto kill_ACKs;
983 }
984
985 if (test_and_clear_bit(RXRPC_CALL_RCVD_ACKALL, &call->events)) {
986 _debug("do implicit ackall");
987 rxrpc_clear_tx_window(call);
988 }
989
990 if (test_bit(RXRPC_CALL_LIFE_TIMER, &call->events)) {
991 write_lock_bh(&call->state_lock);
992 if (call->state <= RXRPC_CALL_COMPLETE) {
993 call->state = RXRPC_CALL_LOCALLY_ABORTED;
994 call->abort_code = RX_CALL_TIMEOUT;
995 set_bit(RXRPC_CALL_ABORT, &call->events);
996 }
997 write_unlock_bh(&call->state_lock);
998
999 _debug("post timeout");
1000 if (rxrpc_post_message(call, RXRPC_SKB_MARK_LOCAL_ERROR,
1001 ETIME, true) < 0)
1002 goto no_mem;
1003
1004 clear_bit(RXRPC_CALL_LIFE_TIMER, &call->events);
1005 goto kill_ACKs;
1006 }
1007
1008 /* deal with assorted inbound messages */
1009 if (!skb_queue_empty(&call->rx_queue)) {
1010 switch (rxrpc_process_rx_queue(call, &abort_code)) {
1011 case 0:
1012 case -EAGAIN:
1013 break;
1014 case -ENOMEM:
1015 goto no_mem;
1016 case -EKEYEXPIRED:
1017 case -EKEYREJECTED:
1018 case -EPROTO:
1019 rxrpc_abort_call(call, abort_code);
1020 goto kill_ACKs;
1021 }
1022 }
1023
1024 /* handle resending */
1025 if (test_and_clear_bit(RXRPC_CALL_RESEND_TIMER, &call->events))
1026 rxrpc_resend_timer(call);
1027 if (test_and_clear_bit(RXRPC_CALL_RESEND, &call->events))
1028 rxrpc_resend(call);
1029
1030 /* consider sending an ordinary ACK */
1031 if (test_bit(RXRPC_CALL_ACK, &call->events)) {
1032 _debug("send ACK: window: %d - %d { %lx }",
1033 call->rx_data_eaten, call->ackr_win_top,
1034 call->ackr_window[0]);
1035
1036 if (call->state > RXRPC_CALL_SERVER_ACK_REQUEST &&
1037 call->ackr_reason != RXRPC_ACK_PING_RESPONSE) {
1038 /* ACK by sending reply DATA packet in this state */
1039 clear_bit(RXRPC_CALL_ACK, &call->events);
1040 goto maybe_reschedule;
1041 }
1042
1043 genbit = RXRPC_CALL_ACK;
1044
1045 acks = kzalloc(call->ackr_win_top - call->rx_data_eaten,
1046 GFP_NOFS);
1047 if (!acks)
1048 goto no_mem;
1049
1050 //hdr.flags = RXRPC_SLOW_START_OK;
1051 ack.bufferSpace = htons(8);
1052 ack.maxSkew = 0;
1053 ack.serial = 0;
1054 ack.reason = 0;
1055
1056 spin_lock_bh(&call->lock);
1057 ack.reason = call->ackr_reason;
1058 ack.serial = call->ackr_serial;
1059 ack.previousPacket = call->ackr_prev_seq;
1060 ack.firstPacket = htonl(call->rx_data_eaten + 1);
1061
1062 ack.nAcks = 0;
1063 for (loop = 0; loop < RXRPC_ACKR_WINDOW_ASZ; loop++) {
1064 nbit = loop * BITS_PER_LONG;
1065 for (bits = call->ackr_window[loop]; bits; bits >>= 1
1066 ) {
1067 _debug("- l=%d n=%d b=%lx", loop, nbit, bits);
1068 if (bits & 1) {
1069 acks[nbit] = RXRPC_ACK_TYPE_ACK;
1070 ack.nAcks = nbit + 1;
1071 }
1072 nbit++;
1073 }
1074 }
1075 call->ackr_reason = 0;
1076 spin_unlock_bh(&call->lock);
1077
1078 pad = 0;
1079
1080 iov[1].iov_base = &ack;
1081 iov[1].iov_len = sizeof(ack);
1082 iov[2].iov_base = acks;
1083 iov[2].iov_len = ack.nAcks;
1084 iov[3].iov_base = &pad;
1085 iov[3].iov_len = 3;
1086 iov[4].iov_base = &ackinfo;
1087 iov[4].iov_len = sizeof(ackinfo);
1088
1089 switch (ack.reason) {
1090 case RXRPC_ACK_REQUESTED:
1091 case RXRPC_ACK_DUPLICATE:
1092 case RXRPC_ACK_OUT_OF_SEQUENCE:
1093 case RXRPC_ACK_EXCEEDS_WINDOW:
1094 case RXRPC_ACK_NOSPACE:
1095 case RXRPC_ACK_PING:
1096 case RXRPC_ACK_PING_RESPONSE:
1097 goto send_ACK_with_skew;
1098 case RXRPC_ACK_DELAY:
1099 case RXRPC_ACK_IDLE:
1100 goto send_ACK;
1101 }
1102 }
1103
1104 /* handle completion of security negotiations on an incoming
1105 * connection */
1106 if (test_and_clear_bit(RXRPC_CALL_SECURED, &call->events)) {
1107 _debug("secured");
1108 spin_lock_bh(&call->lock);
1109
1110 if (call->state == RXRPC_CALL_SERVER_SECURING) {
1111 _debug("securing");
1112 write_lock(&call->conn->lock);
1113 if (!test_bit(RXRPC_CALL_RELEASED, &call->flags) &&
1114 !test_bit(RXRPC_CALL_RELEASE, &call->events)) {
1115 _debug("not released");
1116 call->state = RXRPC_CALL_SERVER_ACCEPTING;
1117 list_move_tail(&call->accept_link,
1118 &call->socket->acceptq);
1119 }
1120 write_unlock(&call->conn->lock);
1121 read_lock(&call->state_lock);
1122 if (call->state < RXRPC_CALL_COMPLETE)
1123 set_bit(RXRPC_CALL_POST_ACCEPT, &call->events);
1124 read_unlock(&call->state_lock);
1125 }
1126
1127 spin_unlock_bh(&call->lock);
1128 if (!test_bit(RXRPC_CALL_POST_ACCEPT, &call->events))
1129 goto maybe_reschedule;
1130 }
1131
1132 /* post a notification of an acceptable connection to the app */
1133 if (test_bit(RXRPC_CALL_POST_ACCEPT, &call->events)) {
1134 _debug("post accept");
1135 if (rxrpc_post_message(call, RXRPC_SKB_MARK_NEW_CALL,
1136 0, false) < 0)
1137 goto no_mem;
1138 clear_bit(RXRPC_CALL_POST_ACCEPT, &call->events);
1139 goto maybe_reschedule;
1140 }
1141
1142 /* handle incoming call acceptance */
1143 if (test_and_clear_bit(RXRPC_CALL_ACCEPTED, &call->events)) {
1144 _debug("accepted");
1145 ASSERTCMP(call->rx_data_post, ==, 0);
1146 call->rx_data_post = 1;
1147 read_lock_bh(&call->state_lock);
1148 if (call->state < RXRPC_CALL_COMPLETE)
1149 set_bit(RXRPC_CALL_DRAIN_RX_OOS, &call->events);
1150 read_unlock_bh(&call->state_lock);
1151 }
1152
1153 /* drain the out of sequence received packet queue into the packet Rx
1154 * queue */
1155 if (test_and_clear_bit(RXRPC_CALL_DRAIN_RX_OOS, &call->events)) {
1156 while (call->rx_data_post == call->rx_first_oos)
1157 if (rxrpc_drain_rx_oos_queue(call) < 0)
1158 break;
1159 goto maybe_reschedule;
1160 }
1161
1162 /* other events may have been raised since we started checking */
1163 goto maybe_reschedule;
1164
1165 send_ACK_with_skew:
1166 ack.maxSkew = htons(atomic_read(&call->conn->hi_serial) -
1167 ntohl(ack.serial));
1168 send_ACK:
1169 mtu = call->conn->trans->peer->if_mtu;
1170 mtu -= call->conn->trans->peer->hdrsize;
1171 ackinfo.maxMTU = htonl(mtu);
1172 ackinfo.rwind = htonl(32);
1173
1174 /* permit the peer to send us jumbo packets if it wants to */
1175 ackinfo.rxMTU = htonl(5692);
1176 ackinfo.jumbo_max = htonl(4);
1177
1178 hdr.serial = htonl(atomic_inc_return(&call->conn->serial));
1179 _proto("Tx ACK %%%u { m=%hu f=#%u p=#%u s=%%%u r=%s n=%u }",
1180 ntohl(hdr.serial),
1181 ntohs(ack.maxSkew),
1182 ntohl(ack.firstPacket),
1183 ntohl(ack.previousPacket),
1184 ntohl(ack.serial),
1185 rxrpc_acks[ack.reason],
1186 ack.nAcks);
1187
1188 del_timer_sync(&call->ack_timer);
1189 if (ack.nAcks > 0)
1190 set_bit(RXRPC_CALL_TX_SOFT_ACK, &call->flags);
1191 goto send_message_2;
1192
1193 send_message:
1194 _debug("send message");
1195
1196 hdr.serial = htonl(atomic_inc_return(&call->conn->serial));
1197 _proto("Tx %s %%%u", rxrpc_pkts[hdr.type], ntohl(hdr.serial));
1198 send_message_2:
1199
1200 len = iov[0].iov_len;
1201 ioc = 1;
1202 if (iov[4].iov_len) {
1203 ioc = 5;
1204 len += iov[4].iov_len;
1205 len += iov[3].iov_len;
1206 len += iov[2].iov_len;
1207 len += iov[1].iov_len;
1208 } else if (iov[3].iov_len) {
1209 ioc = 4;
1210 len += iov[3].iov_len;
1211 len += iov[2].iov_len;
1212 len += iov[1].iov_len;
1213 } else if (iov[2].iov_len) {
1214 ioc = 3;
1215 len += iov[2].iov_len;
1216 len += iov[1].iov_len;
1217 } else if (iov[1].iov_len) {
1218 ioc = 2;
1219 len += iov[1].iov_len;
1220 }
1221
1222 ret = kernel_sendmsg(call->conn->trans->local->socket,
1223 &msg, iov, ioc, len);
1224 if (ret < 0) {
1225 _debug("sendmsg failed: %d", ret);
1226 read_lock_bh(&call->state_lock);
1227 if (call->state < RXRPC_CALL_DEAD)
1228 rxrpc_queue_call(call);
1229 read_unlock_bh(&call->state_lock);
1230 goto error;
1231 }
1232
1233 switch (genbit) {
1234 case RXRPC_CALL_ABORT:
1235 clear_bit(genbit, &call->events);
1236 clear_bit(RXRPC_CALL_RCVD_ABORT, &call->events);
1237 goto kill_ACKs;
1238
1239 case RXRPC_CALL_ACK_FINAL:
1240 write_lock_bh(&call->state_lock);
1241 if (call->state == RXRPC_CALL_CLIENT_FINAL_ACK)
1242 call->state = RXRPC_CALL_COMPLETE;
1243 write_unlock_bh(&call->state_lock);
1244 goto kill_ACKs;
1245
1246 default:
1247 clear_bit(genbit, &call->events);
1248 switch (call->state) {
1249 case RXRPC_CALL_CLIENT_AWAIT_REPLY:
1250 case RXRPC_CALL_CLIENT_RECV_REPLY:
1251 case RXRPC_CALL_SERVER_RECV_REQUEST:
1252 case RXRPC_CALL_SERVER_ACK_REQUEST:
1253 _debug("start ACK timer");
1254 rxrpc_propose_ACK(call, RXRPC_ACK_DELAY,
1255 call->ackr_serial, false);
1256 default:
1257 break;
1258 }
1259 goto maybe_reschedule;
1260 }
1261
1262 kill_ACKs:
1263 del_timer_sync(&call->ack_timer);
1264 if (test_and_clear_bit(RXRPC_CALL_ACK_FINAL, &call->events))
1265 rxrpc_put_call(call);
1266 clear_bit(RXRPC_CALL_ACK, &call->events);
1267
1268 maybe_reschedule:
1269 if (call->events || !skb_queue_empty(&call->rx_queue)) {
1270 read_lock_bh(&call->state_lock);
1271 if (call->state < RXRPC_CALL_DEAD)
1272 rxrpc_queue_call(call);
1273 read_unlock_bh(&call->state_lock);
1274 }
1275
1276 /* don't leave aborted connections on the accept queue */
1277 if (call->state >= RXRPC_CALL_COMPLETE &&
1278 !list_empty(&call->accept_link)) {
1279 _debug("X unlinking once-pending call %p { e=%lx f=%lx c=%x }",
1280 call, call->events, call->flags,
1281 ntohl(call->conn->cid));
1282
1283 read_lock_bh(&call->state_lock);
1284 if (!test_bit(RXRPC_CALL_RELEASED, &call->flags) &&
1285 !test_and_set_bit(RXRPC_CALL_RELEASE, &call->events))
1286 rxrpc_queue_call(call);
1287 read_unlock_bh(&call->state_lock);
1288 }
1289
1290 error:
1291 clear_bit(RXRPC_CALL_PROC_BUSY, &call->flags);
1292 kfree(acks);
1293
1294 /* because we don't want two CPUs both processing the work item for one
1295 * call at the same time, we use a flag to note when it's busy; however
1296 * this means there's a race between clearing the flag and setting the
1297 * work pending bit and the work item being processed again */
1298 if (call->events && !work_pending(&call->processor)) {
1299 _debug("jumpstart %x", ntohl(call->conn->cid));
1300 rxrpc_queue_call(call);
1301 }
1302
1303 _leave("");
1304 return;
1305
1306 no_mem:
1307 _debug("out of memory");
1308 goto maybe_reschedule;
1309 }