1 /*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994, 1995
5 * The Regents of the University of California.
6 * All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the University nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
19 *
20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 * SUCH DAMAGE.
31 */
32
33 /*-
34 *
35 * NRL grants permission for redistribution and use in source and binary
36 * forms, with or without modification, of the software and documentation
37 * created at NRL provided that the following conditions are met:
38 *
39 * 1. Redistributions of source code must retain the above copyright
40 * notice, this list of conditions and the following disclaimer.
41 * 2. Redistributions in binary form must reproduce the above copyright
42 * notice, this list of conditions and the following disclaimer in the
43 * documentation and/or other materials provided with the distribution.
44 * 3. All advertising materials mentioning features or use of this software
45 * must display the following acknowledgements:
46 * This product includes software developed by the University of
47 * California, Berkeley and its contributors.
48 * This product includes software developed at the Information
49 * Technology Division, US Naval Research Laboratory.
50 * 4. Neither the name of the NRL nor the names of its contributors
51 * may be used to endorse or promote products derived from this software
52 * without specific prior written permission.
53 *
54 * THE SOFTWARE PROVIDED BY NRL IS PROVIDED BY NRL AND CONTRIBUTORS ``AS
55 * IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
56 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
57 * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NRL OR
58 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
59 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
60 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
61 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
62 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
63 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
64 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
65 *
66 * The views and conclusions contained in the software and documentation
67 * are those of the authors and should not be interpreted as representing
68 * official policies, either expressed or implied, of the US Naval
69 * Research Laboratory (NRL).
70 */
71
72 #include <sys/cdefs.h>
73 #include "opt_inet.h"
74 #include "opt_inet6.h"
75
76 #include <sys/param.h>
77 #include <sys/systm.h>
78 #include <sys/kernel.h>
79 #include <sys/sysctl.h>
80 #include <sys/malloc.h>
81 #include <sys/mbuf.h>
82 #include <sys/proc.h> /* for proc0 declaration */
83 #include <sys/protosw.h>
84 #include <sys/socket.h>
85 #include <sys/socketvar.h>
86 #include <sys/syslog.h>
87 #include <sys/systm.h>
88
89 #include <machine/cpu.h> /* before tcp_seq.h, for tcp_random18() */
90
91 #include <vm/uma.h>
92
93 #include <net/if.h>
94 #include <net/if_var.h>
95 #include <net/route.h>
96 #include <net/vnet.h>
97
98 #include <netinet/in.h>
99 #include <netinet/in_systm.h>
100 #include <netinet/ip.h>
101 #include <netinet/in_var.h>
102 #include <netinet/in_pcb.h>
103 #include <netinet/ip_var.h>
104 #include <netinet/ip6.h>
105 #include <netinet/icmp6.h>
106 #include <netinet6/nd6.h>
107 #include <netinet6/ip6_var.h>
108 #include <netinet6/in6_pcb.h>
109 #include <netinet/tcp.h>
110 #include <netinet/tcp_fsm.h>
111 #include <netinet/tcp_seq.h>
112 #include <netinet/tcp_timer.h>
113 #include <netinet/tcp_var.h>
114 #include <netinet/tcpip.h>
115 #include <netinet/cc/cc.h>
116
117 #include <machine/in_cksum.h>
118
119 VNET_DECLARE(struct uma_zone *, sack_hole_zone);
120 #define V_sack_hole_zone VNET(sack_hole_zone)
121
122 SYSCTL_NODE(_net_inet_tcp, OID_AUTO, sack, CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
123 "TCP SACK");
124
125 VNET_DEFINE(int, tcp_do_sack) = 1;
126 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, enable, CTLFLAG_VNET | CTLFLAG_RW,
127 &VNET_NAME(tcp_do_sack), 0,
128 "Enable/Disable TCP SACK support");
129
130 VNET_DEFINE(int, tcp_do_newsack) = 1;
131 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, revised, CTLFLAG_VNET | CTLFLAG_RW,
132 &VNET_NAME(tcp_do_newsack), 0,
133 "Use revised SACK loss recovery per RFC 6675");
134
135 VNET_DEFINE(int, tcp_do_lrd) = 1;
136 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, lrd, CTLFLAG_VNET | CTLFLAG_RW,
137 &VNET_NAME(tcp_do_lrd), 1,
138 "Perform Lost Retransmission Detection");
139
140 VNET_DEFINE(int, tcp_sack_tso) = 0;
141 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, tso, CTLFLAG_VNET | CTLFLAG_RW,
142 &VNET_NAME(tcp_sack_tso), 0,
143 "Allow TSO during SACK loss recovery");
144
145 VNET_DEFINE(int, tcp_sack_maxholes) = 128;
146 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, maxholes, CTLFLAG_VNET | CTLFLAG_RW,
147 &VNET_NAME(tcp_sack_maxholes), 0,
148 "Maximum number of TCP SACK holes allowed per connection");
149
150 VNET_DEFINE(int, tcp_sack_globalmaxholes) = 65536;
151 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, globalmaxholes, CTLFLAG_VNET | CTLFLAG_RW,
152 &VNET_NAME(tcp_sack_globalmaxholes), 0,
153 "Global maximum number of TCP SACK holes");
154
155 VNET_DEFINE(int, tcp_sack_globalholes) = 0;
156 SYSCTL_INT(_net_inet_tcp_sack, OID_AUTO, globalholes, CTLFLAG_VNET | CTLFLAG_RD,
157 &VNET_NAME(tcp_sack_globalholes), 0,
158 "Global number of TCP SACK holes currently allocated");
159
160 int
tcp_dsack_block_exists(struct tcpcb * tp)161 tcp_dsack_block_exists(struct tcpcb *tp)
162 {
163 /* Return true if a DSACK block exists */
164 if (tp->rcv_numsacks == 0)
165 return (0);
166 if (SEQ_LEQ(tp->sackblks[0].end, tp->rcv_nxt))
167 return(1);
168 return (0);
169 }
170
171 /*
172 * This function will find overlaps with the currently stored sackblocks
173 * and add any overlap as a dsack block upfront
174 */
175 void
tcp_update_dsack_list(struct tcpcb * tp,tcp_seq rcv_start,tcp_seq rcv_end)176 tcp_update_dsack_list(struct tcpcb *tp, tcp_seq rcv_start, tcp_seq rcv_end)
177 {
178 struct sackblk head_blk,mid_blk,saved_blks[MAX_SACK_BLKS];
179 int i, j, n, identical;
180 tcp_seq start, end;
181
182 INP_WLOCK_ASSERT(tptoinpcb(tp));
183
184 KASSERT(SEQ_LT(rcv_start, rcv_end), ("rcv_start < rcv_end"));
185
186 if (SEQ_LT(rcv_end, tp->rcv_nxt) ||
187 ((rcv_end == tp->rcv_nxt) &&
188 (tp->rcv_numsacks > 0 ) &&
189 (tp->sackblks[0].end == tp->rcv_nxt))) {
190 saved_blks[0].start = rcv_start;
191 saved_blks[0].end = rcv_end;
192 } else {
193 saved_blks[0].start = saved_blks[0].end = 0;
194 }
195
196 head_blk.start = head_blk.end = 0;
197 mid_blk.start = rcv_start;
198 mid_blk.end = rcv_end;
199 identical = 0;
200
201 for (i = 0; i < tp->rcv_numsacks; i++) {
202 start = tp->sackblks[i].start;
203 end = tp->sackblks[i].end;
204 if (SEQ_LT(rcv_end, start)) {
205 /* pkt left to sack blk */
206 continue;
207 }
208 if (SEQ_GT(rcv_start, end)) {
209 /* pkt right to sack blk */
210 continue;
211 }
212 if (SEQ_GT(tp->rcv_nxt, end)) {
213 if ((SEQ_MAX(rcv_start, start) != SEQ_MIN(rcv_end, end)) &&
214 (SEQ_GT(head_blk.start, SEQ_MAX(rcv_start, start)) ||
215 (head_blk.start == head_blk.end))) {
216 head_blk.start = SEQ_MAX(rcv_start, start);
217 head_blk.end = SEQ_MIN(rcv_end, end);
218 }
219 continue;
220 }
221 if (((head_blk.start == head_blk.end) ||
222 SEQ_LT(start, head_blk.start)) &&
223 (SEQ_GT(end, rcv_start) &&
224 SEQ_LEQ(start, rcv_end))) {
225 head_blk.start = start;
226 head_blk.end = end;
227 }
228 mid_blk.start = SEQ_MIN(mid_blk.start, start);
229 mid_blk.end = SEQ_MAX(mid_blk.end, end);
230 if ((mid_blk.start == start) &&
231 (mid_blk.end == end))
232 identical = 1;
233 }
234 if (SEQ_LT(head_blk.start, head_blk.end)) {
235 /* store overlapping range */
236 saved_blks[0].start = SEQ_MAX(rcv_start, head_blk.start);
237 saved_blks[0].end = SEQ_MIN(rcv_end, head_blk.end);
238 }
239 n = 1;
240 /*
241 * Second, if not ACKed, store the SACK block that
242 * overlaps with the DSACK block unless it is identical
243 */
244 if ((SEQ_LT(tp->rcv_nxt, mid_blk.end) &&
245 !((mid_blk.start == saved_blks[0].start) &&
246 (mid_blk.end == saved_blks[0].end))) ||
247 identical == 1) {
248 saved_blks[n].start = mid_blk.start;
249 saved_blks[n++].end = mid_blk.end;
250 }
251 for (j = 0; (j < tp->rcv_numsacks) && (n < MAX_SACK_BLKS); j++) {
252 if (((SEQ_LT(tp->sackblks[j].end, mid_blk.start) ||
253 SEQ_GT(tp->sackblks[j].start, mid_blk.end)) &&
254 (SEQ_GT(tp->sackblks[j].start, tp->rcv_nxt))))
255 saved_blks[n++] = tp->sackblks[j];
256 }
257 j = 0;
258 for (i = 0; i < n; i++) {
259 /* we can end up with a stale initial entry */
260 if (SEQ_LT(saved_blks[i].start, saved_blks[i].end)) {
261 tp->sackblks[j++] = saved_blks[i];
262 }
263 }
264 tp->rcv_numsacks = j;
265 }
266
267 /*
268 * This function is called upon receipt of new valid data (while not in
269 * header prediction mode), and it updates the ordered list of sacks.
270 */
271 void
tcp_update_sack_list(struct tcpcb * tp,tcp_seq rcv_start,tcp_seq rcv_end)272 tcp_update_sack_list(struct tcpcb *tp, tcp_seq rcv_start, tcp_seq rcv_end)
273 {
274 /*
275 * First reported block MUST be the most recent one. Subsequent
276 * blocks SHOULD be in the order in which they arrived at the
277 * receiver. These two conditions make the implementation fully
278 * compliant with RFC 2018.
279 */
280 struct sackblk head_blk, saved_blks[MAX_SACK_BLKS];
281 int num_head, num_saved, i;
282
283 INP_WLOCK_ASSERT(tptoinpcb(tp));
284
285 /* Check arguments. */
286 KASSERT(SEQ_LEQ(rcv_start, rcv_end), ("rcv_start <= rcv_end"));
287
288 if ((rcv_start == rcv_end) &&
289 (tp->rcv_numsacks >= 1) &&
290 (rcv_end == tp->sackblks[0].end)) {
291 /* retaining DSACK block below rcv_nxt (todrop) */
292 head_blk = tp->sackblks[0];
293 } else {
294 /* SACK block for the received segment. */
295 head_blk.start = rcv_start;
296 head_blk.end = rcv_end;
297 }
298
299 /*
300 * Merge updated SACK blocks into head_blk, and save unchanged SACK
301 * blocks into saved_blks[]. num_saved will have the number of the
302 * saved SACK blocks.
303 */
304 num_saved = 0;
305 for (i = 0; i < tp->rcv_numsacks; i++) {
306 tcp_seq start = tp->sackblks[i].start;
307 tcp_seq end = tp->sackblks[i].end;
308 if (SEQ_GEQ(start, end) || SEQ_LEQ(start, tp->rcv_nxt)) {
309 /*
310 * Discard this SACK block.
311 */
312 } else if (SEQ_LEQ(head_blk.start, end) &&
313 SEQ_GEQ(head_blk.end, start)) {
314 /*
315 * Merge this SACK block into head_blk. This SACK
316 * block itself will be discarded.
317 */
318 /*
319 * |-|
320 * |---| merge
321 *
322 * |-|
323 * |---| merge
324 *
325 * |-----|
326 * |-| DSACK smaller
327 *
328 * |-|
329 * |-----| DSACK smaller
330 */
331 if (head_blk.start == end)
332 head_blk.start = start;
333 else if (head_blk.end == start)
334 head_blk.end = end;
335 else {
336 if (SEQ_LT(head_blk.start, start)) {
337 tcp_seq temp = start;
338 start = head_blk.start;
339 head_blk.start = temp;
340 }
341 if (SEQ_GT(head_blk.end, end)) {
342 tcp_seq temp = end;
343 end = head_blk.end;
344 head_blk.end = temp;
345 }
346 if ((head_blk.start != start) ||
347 (head_blk.end != end)) {
348 if ((num_saved >= 1) &&
349 SEQ_GEQ(saved_blks[num_saved-1].start, start) &&
350 SEQ_LEQ(saved_blks[num_saved-1].end, end))
351 num_saved--;
352 saved_blks[num_saved].start = start;
353 saved_blks[num_saved].end = end;
354 num_saved++;
355 }
356 }
357 } else {
358 /*
359 * This block supercedes the prior block
360 */
361 if ((num_saved >= 1) &&
362 SEQ_GEQ(saved_blks[num_saved-1].start, start) &&
363 SEQ_LEQ(saved_blks[num_saved-1].end, end))
364 num_saved--;
365 /*
366 * Save this SACK block.
367 */
368 saved_blks[num_saved].start = start;
369 saved_blks[num_saved].end = end;
370 num_saved++;
371 }
372 }
373
374 /*
375 * Update SACK list in tp->sackblks[].
376 */
377 num_head = 0;
378 if (SEQ_LT(rcv_start, rcv_end)) {
379 /*
380 * The received data segment is an out-of-order segment. Put
381 * head_blk at the top of SACK list.
382 */
383 tp->sackblks[0] = head_blk;
384 num_head = 1;
385 /*
386 * If the number of saved SACK blocks exceeds its limit,
387 * discard the last SACK block.
388 */
389 if (num_saved >= MAX_SACK_BLKS)
390 num_saved--;
391 }
392 if ((rcv_start == rcv_end) &&
393 (rcv_start == tp->sackblks[0].end)) {
394 num_head = 1;
395 }
396 if (num_saved > 0) {
397 /*
398 * Copy the saved SACK blocks back.
399 */
400 bcopy(saved_blks, &tp->sackblks[num_head],
401 sizeof(struct sackblk) * num_saved);
402 }
403
404 /* Save the number of SACK blocks. */
405 tp->rcv_numsacks = num_head + num_saved;
406 }
407
408 void
tcp_clean_dsack_blocks(struct tcpcb * tp)409 tcp_clean_dsack_blocks(struct tcpcb *tp)
410 {
411 struct sackblk saved_blks[MAX_SACK_BLKS];
412 int num_saved, i;
413
414 INP_WLOCK_ASSERT(tptoinpcb(tp));
415 /*
416 * Clean up any DSACK blocks that
417 * are in our queue of sack blocks.
418 *
419 */
420 num_saved = 0;
421 for (i = 0; i < tp->rcv_numsacks; i++) {
422 tcp_seq start = tp->sackblks[i].start;
423 tcp_seq end = tp->sackblks[i].end;
424 if (SEQ_GEQ(start, end) || SEQ_LEQ(start, tp->rcv_nxt)) {
425 /*
426 * Discard this D-SACK block.
427 */
428 continue;
429 }
430 /*
431 * Save this SACK block.
432 */
433 saved_blks[num_saved].start = start;
434 saved_blks[num_saved].end = end;
435 num_saved++;
436 }
437 if (num_saved > 0) {
438 /*
439 * Copy the saved SACK blocks back.
440 */
441 bcopy(saved_blks, &tp->sackblks[0],
442 sizeof(struct sackblk) * num_saved);
443 }
444 tp->rcv_numsacks = num_saved;
445 }
446
447 /*
448 * Delete all receiver-side SACK information.
449 */
450 void
tcp_clean_sackreport(struct tcpcb * tp)451 tcp_clean_sackreport(struct tcpcb *tp)
452 {
453 int i;
454
455 INP_WLOCK_ASSERT(tptoinpcb(tp));
456 tp->rcv_numsacks = 0;
457 for (i = 0; i < MAX_SACK_BLKS; i++)
458 tp->sackblks[i].start = tp->sackblks[i].end=0;
459 }
460
461 /*
462 * Allocate struct sackhole.
463 */
464 static struct sackhole *
tcp_sackhole_alloc(struct tcpcb * tp,tcp_seq start,tcp_seq end)465 tcp_sackhole_alloc(struct tcpcb *tp, tcp_seq start, tcp_seq end)
466 {
467 struct sackhole *hole;
468
469 if (tp->snd_numholes >= V_tcp_sack_maxholes ||
470 V_tcp_sack_globalholes >= V_tcp_sack_globalmaxholes) {
471 TCPSTAT_INC(tcps_sack_sboverflow);
472 return NULL;
473 }
474
475 hole = (struct sackhole *)uma_zalloc(V_sack_hole_zone, M_NOWAIT);
476 if (hole == NULL)
477 return NULL;
478
479 hole->start = start;
480 hole->end = end;
481 hole->rxmit = start;
482
483 tp->snd_numholes++;
484 atomic_add_int(&V_tcp_sack_globalholes, 1);
485
486 return hole;
487 }
488
489 /*
490 * Free struct sackhole.
491 */
492 static void
tcp_sackhole_free(struct tcpcb * tp,struct sackhole * hole)493 tcp_sackhole_free(struct tcpcb *tp, struct sackhole *hole)
494 {
495
496 uma_zfree(V_sack_hole_zone, hole);
497
498 tp->snd_numholes--;
499 atomic_subtract_int(&V_tcp_sack_globalholes, 1);
500
501 KASSERT(tp->snd_numholes >= 0, ("tp->snd_numholes >= 0"));
502 KASSERT(V_tcp_sack_globalholes >= 0, ("tcp_sack_globalholes >= 0"));
503 }
504
505 /*
506 * Insert new SACK hole into scoreboard.
507 */
508 static struct sackhole *
tcp_sackhole_insert(struct tcpcb * tp,tcp_seq start,tcp_seq end,struct sackhole * after)509 tcp_sackhole_insert(struct tcpcb *tp, tcp_seq start, tcp_seq end,
510 struct sackhole *after)
511 {
512 struct sackhole *hole;
513
514 /* Allocate a new SACK hole. */
515 hole = tcp_sackhole_alloc(tp, start, end);
516 if (hole == NULL)
517 return NULL;
518
519 /* Insert the new SACK hole into scoreboard. */
520 if (after != NULL)
521 TAILQ_INSERT_AFTER(&tp->snd_holes, after, hole, scblink);
522 else
523 TAILQ_INSERT_TAIL(&tp->snd_holes, hole, scblink);
524
525 /* Update SACK hint. */
526 if (tp->sackhint.nexthole == NULL)
527 tp->sackhint.nexthole = hole;
528
529 return hole;
530 }
531
532 /*
533 * Remove SACK hole from scoreboard.
534 */
535 static void
tcp_sackhole_remove(struct tcpcb * tp,struct sackhole * hole)536 tcp_sackhole_remove(struct tcpcb *tp, struct sackhole *hole)
537 {
538
539 /* Update SACK hint. */
540 if (tp->sackhint.nexthole == hole)
541 tp->sackhint.nexthole = TAILQ_NEXT(hole, scblink);
542
543 /* Remove this SACK hole. */
544 TAILQ_REMOVE(&tp->snd_holes, hole, scblink);
545
546 /* Free this SACK hole. */
547 tcp_sackhole_free(tp, hole);
548 }
549
550 /*
551 * Process cumulative ACK and the TCP SACK option to update the scoreboard.
552 * tp->snd_holes is an ordered list of holes (oldest to newest, in terms of
553 * the sequence space).
554 * Returns SACK_NEWLOSS if incoming ACK indicates ongoing loss (hole split, new hole),
555 * SACK_CHANGE if incoming ACK has previously unknown SACK information,
556 * SACK_NOCHANGE otherwise.
557 */
558 sackstatus_t
tcp_sack_doack(struct tcpcb * tp,struct tcpopt * to,tcp_seq th_ack)559 tcp_sack_doack(struct tcpcb *tp, struct tcpopt *to, tcp_seq th_ack)
560 {
561 struct sackhole *cur, *temp;
562 struct sackblk sack, sack_blocks[TCP_MAX_SACK + 1], *sblkp;
563 int i, j, num_sack_blks;
564 sackstatus_t sack_changed;
565 int delivered_data, left_edge_delta;
566 int maxseg = tp->t_maxseg - MAX_TCPOPTLEN;
567
568 tcp_seq loss_hiack = 0;
569 int loss_thresh = 0;
570 int loss_sblks = 0;
571 int notlost_bytes = 0;
572
573 INP_WLOCK_ASSERT(tptoinpcb(tp));
574
575 num_sack_blks = 0;
576 sack_changed = SACK_NOCHANGE;
577 delivered_data = 0;
578 left_edge_delta = 0;
579 /*
580 * If SND.UNA will be advanced by SEG.ACK, and if SACK holes exist,
581 * treat [SND.UNA, SEG.ACK) as if it is a SACK block.
582 * Account changes to SND.UNA always in delivered data.
583 */
584 if (SEQ_LT(tp->snd_una, th_ack) && !TAILQ_EMPTY(&tp->snd_holes)) {
585 left_edge_delta = th_ack - tp->snd_una;
586 sack_blocks[num_sack_blks].start = tp->snd_una;
587 sack_blocks[num_sack_blks++].end = th_ack;
588 /*
589 * Pulling snd_fack forward if we got here
590 * due to DSACK blocks
591 */
592 if (SEQ_LT(tp->snd_fack, th_ack)) {
593 delivered_data += th_ack - tp->snd_una;
594 tp->snd_fack = th_ack;
595 sack_changed = SACK_CHANGE;
596 }
597 }
598 /*
599 * Append received valid SACK blocks to sack_blocks[], but only if we
600 * received new blocks from the other side.
601 */
602 if (to->to_flags & TOF_SACK) {
603 for (i = 0; i < to->to_nsacks; i++) {
604 bcopy((to->to_sacks + i * TCPOLEN_SACK),
605 &sack, sizeof(sack));
606 sack.start = ntohl(sack.start);
607 sack.end = ntohl(sack.end);
608 if (SEQ_GT(sack.end, sack.start) &&
609 SEQ_GT(sack.start, tp->snd_una) &&
610 SEQ_GT(sack.start, th_ack) &&
611 SEQ_LT(sack.start, tp->snd_max) &&
612 SEQ_GT(sack.end, tp->snd_una) &&
613 SEQ_LEQ(sack.end, tp->snd_max) &&
614 ((sack.end - sack.start) >= maxseg ||
615 SEQ_GEQ(sack.end, tp->snd_max))) {
616 sack_blocks[num_sack_blks++] = sack;
617 } else if (SEQ_LEQ(sack.start, th_ack) &&
618 SEQ_LEQ(sack.end, th_ack)) {
619 /*
620 * Its a D-SACK block.
621 */
622 tcp_record_dsack(tp, sack.start, sack.end, 0);
623 }
624 }
625 }
626 /*
627 * Return if SND.UNA is not advanced and no valid SACK block is
628 * received.
629 */
630 if (num_sack_blks == 0)
631 return (sack_changed);
632
633 /*
634 * Sort the SACK blocks so we can update the scoreboard with just one
635 * pass. The overhead of sorting up to 4+1 elements is less than
636 * making up to 4+1 passes over the scoreboard.
637 */
638 for (i = 0; i < num_sack_blks; i++) {
639 for (j = i + 1; j < num_sack_blks; j++) {
640 if (SEQ_GT(sack_blocks[i].end, sack_blocks[j].end)) {
641 sack = sack_blocks[i];
642 sack_blocks[i] = sack_blocks[j];
643 sack_blocks[j] = sack;
644 }
645 }
646 }
647 if (TAILQ_EMPTY(&tp->snd_holes)) {
648 /*
649 * Empty scoreboard. Need to initialize snd_fack (it may be
650 * uninitialized or have a bogus value). Scoreboard holes
651 * (from the sack blocks received) are created later below
652 * (in the logic that adds holes to the tail of the
653 * scoreboard).
654 */
655 tp->snd_fack = SEQ_MAX(tp->snd_una, th_ack);
656 tp->sackhint.sacked_bytes = 0; /* reset */
657 tp->sackhint.hole_bytes = 0;
658 }
659 /*
660 * In the while-loop below, incoming SACK blocks (sack_blocks[]) and
661 * SACK holes (snd_holes) are traversed from their tails with just
662 * one pass in order to reduce the number of compares especially when
663 * the bandwidth-delay product is large.
664 *
665 * Note: Typically, in the first RTT of SACK recovery, the highest
666 * three or four SACK blocks with the same ack number are received.
667 * In the second RTT, if retransmitted data segments are not lost,
668 * the highest three or four SACK blocks with ack number advancing
669 * are received.
670 */
671 sblkp = &sack_blocks[num_sack_blks - 1]; /* Last SACK block */
672 tp->sackhint.last_sack_ack = sblkp->end;
673 if (SEQ_LT(tp->snd_fack, sblkp->start)) {
674 /*
675 * The highest SACK block is beyond fack. First,
676 * check if there was a successful Rescue Retransmission,
677 * and move this hole left. With normal holes, snd_fack
678 * is always to the right of the end.
679 */
680 if (((temp = TAILQ_LAST(&tp->snd_holes, sackhole_head)) != NULL) &&
681 SEQ_LEQ(tp->snd_fack,temp->end)) {
682 tp->sackhint.hole_bytes -= temp->end - temp->start;
683 temp->start = SEQ_MAX(tp->snd_fack, SEQ_MAX(tp->snd_una, th_ack));
684 temp->end = sblkp->start;
685 temp->rxmit = temp->start;
686 delivered_data += sblkp->end - sblkp->start;
687 tp->sackhint.hole_bytes += temp->end - temp->start;
688 KASSERT(tp->sackhint.hole_bytes >= 0,
689 ("sackhint hole bytes >= 0"));
690 tp->snd_fack = sblkp->end;
691 sblkp--;
692 sack_changed = SACK_NEWLOSS;
693 } else {
694 /*
695 * Append a new SACK hole at the tail. If the
696 * second or later highest SACK blocks are also
697 * beyond the current fack, they will be inserted
698 * by way of hole splitting in the while-loop below.
699 */
700 temp = tcp_sackhole_insert(tp, tp->snd_fack,sblkp->start,NULL);
701 if (temp != NULL) {
702 delivered_data += sblkp->end - sblkp->start;
703 tp->sackhint.hole_bytes += temp->end - temp->start;
704 tp->snd_fack = sblkp->end;
705 /* Go to the previous sack block. */
706 sblkp--;
707 sack_changed = SACK_CHANGE;
708 } else {
709 /*
710 * We failed to add a new hole based on the current
711 * sack block. Skip over all the sack blocks that
712 * fall completely to the right of snd_fack and
713 * proceed to trim the scoreboard based on the
714 * remaining sack blocks. This also trims the
715 * scoreboard for th_ack (which is sack_blocks[0]).
716 */
717 while (sblkp >= sack_blocks &&
718 SEQ_LT(tp->snd_fack, sblkp->start))
719 sblkp--;
720 if (sblkp >= sack_blocks &&
721 SEQ_LT(tp->snd_fack, sblkp->end)) {
722 delivered_data += sblkp->end - tp->snd_fack;
723 tp->snd_fack = sblkp->end;
724 /*
725 * While the Scoreboard didn't change in
726 * size, we only ended up here because
727 * some SACK data had to be dismissed.
728 */
729 sack_changed = SACK_NEWLOSS;
730 }
731 }
732 }
733 } else if (SEQ_LT(tp->snd_fack, sblkp->end)) {
734 /* fack is advanced. */
735 delivered_data += sblkp->end - tp->snd_fack;
736 tp->snd_fack = sblkp->end;
737 sack_changed = SACK_CHANGE;
738 }
739 cur = TAILQ_LAST(&tp->snd_holes, sackhole_head); /* Last SACK hole. */
740 loss_hiack = tp->snd_fack;
741
742 /*
743 * Since the incoming sack blocks are sorted, we can process them
744 * making one sweep of the scoreboard.
745 */
746 while (cur != NULL) {
747 if (!(sblkp >= sack_blocks)) {
748 if (((loss_sblks >= tcprexmtthresh) ||
749 (loss_thresh > (tcprexmtthresh-1)*tp->t_maxseg)))
750 break;
751 loss_thresh += loss_hiack - cur->end;
752 loss_hiack = cur->start;
753 loss_sblks++;
754 if (!((loss_sblks >= tcprexmtthresh) ||
755 (loss_thresh > (tcprexmtthresh-1)*tp->t_maxseg))) {
756 notlost_bytes += cur->end - cur->start;
757 } else {
758 break;
759 }
760 cur = TAILQ_PREV(cur, sackhole_head, scblink);
761 continue;
762 }
763 if (SEQ_GEQ(sblkp->start, cur->end)) {
764 /*
765 * SACKs data beyond the current hole. Go to the
766 * previous sack block.
767 */
768 sblkp--;
769 continue;
770 }
771 if (SEQ_LEQ(sblkp->end, cur->start)) {
772 /*
773 * SACKs data before the current hole. Go to the
774 * previous hole.
775 */
776 loss_thresh += loss_hiack - cur->end;
777 loss_hiack = cur->start;
778 loss_sblks++;
779 if (!((loss_sblks >= tcprexmtthresh) ||
780 (loss_thresh > (tcprexmtthresh-1)*tp->t_maxseg)))
781 notlost_bytes += cur->end - cur->start;
782 cur = TAILQ_PREV(cur, sackhole_head, scblink);
783 continue;
784 }
785 tp->sackhint.sack_bytes_rexmit -=
786 (SEQ_MIN(cur->rxmit, cur->end) - cur->start);
787 KASSERT(tp->sackhint.sack_bytes_rexmit >= 0,
788 ("sackhint bytes rtx >= 0"));
789 sack_changed = SACK_CHANGE;
790 if (SEQ_LEQ(sblkp->start, cur->start)) {
791 /* Data acks at least the beginning of hole. */
792 if (SEQ_GEQ(sblkp->end, cur->end)) {
793 /* Acks entire hole, so delete hole. */
794 delivered_data += (cur->end - cur->start);
795 temp = cur;
796 cur = TAILQ_PREV(cur, sackhole_head, scblink);
797 tp->sackhint.hole_bytes -= temp->end - temp->start;
798 tcp_sackhole_remove(tp, temp);
799 /*
800 * The sack block may ack all or part of the
801 * next hole too, so continue onto the next
802 * hole.
803 */
804 continue;
805 } else {
806 /* Move start of hole forward. */
807 delivered_data += (sblkp->end - cur->start);
808 tp->sackhint.hole_bytes -= sblkp->end - cur->start;
809 cur->start = sblkp->end;
810 cur->rxmit = SEQ_MAX(cur->rxmit, cur->start);
811 }
812 } else {
813 /* Data acks at least the end of hole. */
814 if (SEQ_GEQ(sblkp->end, cur->end)) {
815 /* Move end of hole backward. */
816 delivered_data += (cur->end - sblkp->start);
817 tp->sackhint.hole_bytes -= cur->end - sblkp->start;
818 cur->end = sblkp->start;
819 cur->rxmit = SEQ_MIN(cur->rxmit, cur->end);
820 if ((tp->t_flags & TF_LRD) && SEQ_GEQ(cur->rxmit, cur->end))
821 cur->rxmit = tp->snd_recover;
822 } else {
823 /*
824 * ACKs some data in middle of a hole; need
825 * to split current hole
826 */
827 temp = tcp_sackhole_insert(tp, sblkp->end,
828 cur->end, cur);
829 sack_changed = SACK_NEWLOSS;
830 if (temp != NULL) {
831 if (SEQ_GT(cur->rxmit, temp->rxmit)) {
832 temp->rxmit = cur->rxmit;
833 tp->sackhint.sack_bytes_rexmit +=
834 (SEQ_MIN(temp->rxmit,
835 temp->end) - temp->start);
836 }
837 tp->sackhint.hole_bytes -= sblkp->end - sblkp->start;
838 loss_thresh += loss_hiack - temp->end;
839 loss_hiack = temp->start;
840 loss_sblks++;
841 if (!((loss_sblks >= tcprexmtthresh) ||
842 (loss_thresh > (tcprexmtthresh-1)*tp->t_maxseg)))
843 notlost_bytes += temp->end - temp->start;
844 cur->end = sblkp->start;
845 cur->rxmit = SEQ_MIN(cur->rxmit,
846 cur->end);
847 if ((tp->t_flags & TF_LRD) && SEQ_GEQ(cur->rxmit, cur->end))
848 cur->rxmit = tp->snd_recover;
849 delivered_data += (sblkp->end - sblkp->start);
850 }
851 }
852 }
853 tp->sackhint.sack_bytes_rexmit +=
854 (SEQ_MIN(cur->rxmit, cur->end) - cur->start);
855 /*
856 * Testing sblkp->start against cur->start tells us whether
857 * we're done with the sack block or the sack hole.
858 * Accordingly, we advance one or the other.
859 */
860 if (SEQ_LEQ(sblkp->start, cur->start)) {
861 loss_thresh += loss_hiack - cur->end;
862 loss_hiack = cur->start;
863 loss_sblks++;
864 if (!((loss_sblks >= tcprexmtthresh) ||
865 (loss_thresh > (tcprexmtthresh-1)*tp->t_maxseg)))
866 notlost_bytes += cur->end - cur->start;
867 cur = TAILQ_PREV(cur, sackhole_head, scblink);
868 } else {
869 sblkp--;
870 }
871 }
872
873 KASSERT(!(TAILQ_EMPTY(&tp->snd_holes) && (tp->sackhint.hole_bytes != 0)),
874 ("SACK scoreboard empty, but accounting non-zero\n"));
875
876 KASSERT(notlost_bytes <= tp->sackhint.hole_bytes,
877 ("SACK: more bytes marked notlost than in scoreboard holes"));
878
879 if (!(to->to_flags & TOF_SACK))
880 /*
881 * If this ACK did not contain any
882 * SACK blocks, any only moved the
883 * left edge right, it is a pure
884 * cumulative ACK. Do not count
885 * DupAck for this. Also required
886 * for RFC6675 rescue retransmission.
887 */
888 sack_changed = SACK_NOCHANGE;
889 tp->sackhint.delivered_data = delivered_data;
890 tp->sackhint.sacked_bytes += delivered_data - left_edge_delta;
891 tp->sackhint.lost_bytes = tp->sackhint.hole_bytes - notlost_bytes;
892 KASSERT((delivered_data >= 0), ("delivered_data < 0"));
893 KASSERT((tp->sackhint.sacked_bytes >= 0), ("sacked_bytes < 0"));
894 return (sack_changed);
895 }
896
897 /*
898 * Free all SACK holes to clear the scoreboard.
899 */
900 void
tcp_free_sackholes(struct tcpcb * tp)901 tcp_free_sackholes(struct tcpcb *tp)
902 {
903 struct sackhole *q;
904
905 INP_WLOCK_ASSERT(tptoinpcb(tp));
906 while ((q = TAILQ_FIRST(&tp->snd_holes)) != NULL)
907 tcp_sackhole_remove(tp, q);
908 tp->sackhint.sack_bytes_rexmit = 0;
909 tp->sackhint.delivered_data = 0;
910 tp->sackhint.sacked_bytes = 0;
911 tp->sackhint.hole_bytes = 0;
912 tp->sackhint.lost_bytes = 0;
913
914 KASSERT(tp->snd_numholes == 0, ("tp->snd_numholes == 0"));
915 KASSERT(tp->sackhint.nexthole == NULL,
916 ("tp->sackhint.nexthole == NULL"));
917 }
918
919 /*
920 * Resend all the currently existing SACK holes of
921 * the scoreboard. This is in line with the Errata to
922 * RFC 2018, which allows the use of SACK data past
923 * an RTO to good effect typically.
924 */
925 void
tcp_resend_sackholes(struct tcpcb * tp)926 tcp_resend_sackholes(struct tcpcb *tp)
927 {
928 struct sackhole *p;
929
930 INP_WLOCK_ASSERT(tptoinpcb(tp));
931 TAILQ_FOREACH(p, &tp->snd_holes, scblink) {
932 p->rxmit = p->start;
933 }
934 tp->sackhint.nexthole = TAILQ_FIRST(&tp->snd_holes);
935 tp->sackhint.sack_bytes_rexmit = 0;
936 }
937
938 /*
939 * Partial ack handling within a sack recovery episode. Keeping this very
940 * simple for now. When a partial ack is received, force snd_cwnd to a value
941 * that will allow the sender to transmit no more than 2 segments. If
942 * necessary, a better scheme can be adopted at a later point, but for now,
943 * the goal is to prevent the sender from bursting a large amount of data in
944 * the midst of sack recovery.
945 */
946 void
tcp_sack_partialack(struct tcpcb * tp,struct tcphdr * th,u_int * maxsegp)947 tcp_sack_partialack(struct tcpcb *tp, struct tcphdr *th, u_int *maxsegp)
948 {
949 struct sackhole *temp;
950 int num_segs = 1;
951 u_int maxseg;
952
953 INP_WLOCK_ASSERT(tptoinpcb(tp));
954
955 if (*maxsegp == 0) {
956 *maxsegp = tcp_maxseg(tp);
957 }
958 maxseg = *maxsegp;
959 tcp_timer_activate(tp, TT_REXMT, 0);
960 tp->t_rtttime = 0;
961 /* Send one or 2 segments based on how much new data was acked. */
962 if ((BYTES_THIS_ACK(tp, th) / maxseg) >= 2)
963 num_segs = 2;
964 if (tp->snd_nxt == tp->snd_max) {
965 tp->snd_cwnd = (tp->sackhint.sack_bytes_rexmit +
966 (tp->snd_nxt - tp->snd_recover) + num_segs * maxseg);
967 } else {
968 /*
969 * Since cwnd is not the expected flightsize during
970 * SACK LR, not deflating cwnd allows the partial
971 * ACKed amount to be sent.
972 */
973 }
974 if (tp->snd_cwnd > tp->snd_ssthresh)
975 tp->snd_cwnd = tp->snd_ssthresh;
976 tp->t_flags |= TF_ACKNOW;
977 /*
978 * RFC6675 rescue retransmission
979 * Add a hole between th_ack (snd_una is not yet set) and snd_max,
980 * if this was a pure cumulative ACK and no data was send beyond
981 * recovery point. Since the data in the socket has not been freed
982 * at this point, we check if the scoreboard is empty, and the ACK
983 * delivered some new data, indicating a full ACK. Also, if the
984 * recovery point is still at snd_max, we are probably application
985 * limited. However, this inference might not always be true. The
986 * rescue retransmission may rarely be slightly premature
987 * compared to RFC6675.
988 * The corresponding ACK+SACK will cause any further outstanding
989 * segments to be retransmitted. This addresses a corner case, when
990 * the trailing packets of a window are lost and no further data
991 * is available for sending.
992 */
993 if ((V_tcp_do_newsack) &&
994 SEQ_LT(th->th_ack, tp->snd_recover) &&
995 TAILQ_EMPTY(&tp->snd_holes) &&
996 (tp->sackhint.delivered_data > 0)) {
997 /*
998 * Exclude FIN sequence space in
999 * the hole for the rescue retransmission,
1000 * and also don't create a hole, if only
1001 * the ACK for a FIN is outstanding.
1002 */
1003 tcp_seq highdata = tp->snd_max;
1004 if (tp->t_flags & TF_SENTFIN)
1005 highdata--;
1006 highdata = SEQ_MIN(highdata, tp->snd_recover);
1007 if (SEQ_LT(th->th_ack, highdata)) {
1008 tp->snd_fack = SEQ_MAX(th->th_ack, tp->snd_fack);
1009 if ((temp = tcp_sackhole_insert(tp, SEQ_MAX(th->th_ack,
1010 highdata - maxseg), highdata, NULL)) != NULL) {
1011 tp->sackhint.hole_bytes +=
1012 temp->end - temp->start;
1013 }
1014 }
1015 }
1016 (void) tcp_output(tp);
1017 }
1018
1019 /*
1020 * Returns the next hole to retransmit and the number of retransmitted bytes
1021 * from the scoreboard. We store both the next hole and the number of
1022 * retransmitted bytes as hints (and recompute these on the fly upon SACK/ACK
1023 * reception). This avoids scoreboard traversals completely.
1024 *
1025 * The loop here will traverse *at most* one link. Here's the argument. For
1026 * the loop to traverse more than 1 link before finding the next hole to
1027 * retransmit, we would need to have at least 1 node following the current
1028 * hint with (rxmit == end). But, for all holes following the current hint,
1029 * (start == rxmit), since we have not yet retransmitted from them.
1030 * Therefore, in order to traverse more 1 link in the loop below, we need to
1031 * have at least one node following the current hint with (start == rxmit ==
1032 * end). But that can't happen, (start == end) means that all the data in
1033 * that hole has been sacked, in which case, the hole would have been removed
1034 * from the scoreboard.
1035 */
1036 struct sackhole *
tcp_sack_output(struct tcpcb * tp,int * sack_bytes_rexmt)1037 tcp_sack_output(struct tcpcb *tp, int *sack_bytes_rexmt)
1038 {
1039 struct sackhole *hole = NULL;
1040
1041 INP_WLOCK_ASSERT(tptoinpcb(tp));
1042 *sack_bytes_rexmt = tp->sackhint.sack_bytes_rexmit;
1043 hole = tp->sackhint.nexthole;
1044 if (hole == NULL)
1045 return (hole);
1046 if (SEQ_GEQ(hole->rxmit, hole->end)) {
1047 for (;;) {
1048 hole = TAILQ_NEXT(hole, scblink);
1049 if (hole == NULL)
1050 return (hole);
1051 if (SEQ_LT(hole->rxmit, hole->end)) {
1052 tp->sackhint.nexthole = hole;
1053 break;
1054 }
1055 }
1056 }
1057 KASSERT(SEQ_LT(hole->start, hole->end), ("%s: hole.start >= hole.end", __func__));
1058 if (!(V_tcp_do_newsack)) {
1059 KASSERT(SEQ_LT(hole->start, tp->snd_fack), ("%s: hole.start >= snd.fack", __func__));
1060 KASSERT(SEQ_LT(hole->end, tp->snd_fack), ("%s: hole.end >= snd.fack", __func__));
1061 KASSERT(SEQ_LT(hole->rxmit, tp->snd_fack), ("%s: hole.rxmit >= snd.fack", __func__));
1062 if (SEQ_GEQ(hole->start, hole->end) ||
1063 SEQ_GEQ(hole->start, tp->snd_fack) ||
1064 SEQ_GEQ(hole->end, tp->snd_fack) ||
1065 SEQ_GEQ(hole->rxmit, tp->snd_fack)) {
1066 log(LOG_CRIT,"tcp: invalid SACK hole (%u-%u,%u) vs fwd ack %u, ignoring.\n",
1067 hole->start, hole->end, hole->rxmit, tp->snd_fack);
1068 return (NULL);
1069 }
1070 }
1071 return (hole);
1072 }
1073
1074 /*
1075 * After a timeout, the SACK list may be rebuilt. This SACK information
1076 * should be used to avoid retransmitting SACKed data. This function
1077 * traverses the SACK list to see if snd_nxt should be moved forward.
1078 * In addition, cwnd will be inflated by the sacked bytes traversed when
1079 * moving snd_nxt forward. This prevents a traffic burst after the final
1080 * full ACK, and also keeps ACKs coming back.
1081 */
1082 int
tcp_sack_adjust(struct tcpcb * tp)1083 tcp_sack_adjust(struct tcpcb *tp)
1084 {
1085 int sacked = 0;
1086 struct sackhole *p, *cur = TAILQ_FIRST(&tp->snd_holes);
1087
1088 INP_WLOCK_ASSERT(tptoinpcb(tp));
1089 if (cur == NULL) {
1090 /* No holes */
1091 return (0);
1092 }
1093 if (SEQ_GEQ(tp->snd_nxt, tp->snd_fack)) {
1094 /* We're already beyond any SACKed blocks */
1095 return (tp->sackhint.sacked_bytes);
1096 }
1097 /*
1098 * Two cases for which we want to advance snd_nxt:
1099 * i) snd_nxt lies between end of one hole and beginning of another
1100 * ii) snd_nxt lies between end of last hole and snd_fack
1101 */
1102 while ((p = TAILQ_NEXT(cur, scblink)) != NULL) {
1103 if (SEQ_LT(tp->snd_nxt, cur->end)) {
1104 return (sacked);
1105 }
1106 sacked += p->start - cur->end;
1107 if (SEQ_GEQ(tp->snd_nxt, p->start)) {
1108 cur = p;
1109 } else {
1110 tp->snd_nxt = p->start;
1111 return (sacked);
1112 }
1113 }
1114 if (SEQ_LT(tp->snd_nxt, cur->end)) {
1115 return (sacked);
1116 }
1117 tp->snd_nxt = tp->snd_fack;
1118 return (tp->sackhint.sacked_bytes);
1119 }
1120
1121 /*
1122 * Lost Retransmission Detection
1123 * Check is FACK is beyond the rexmit of the leftmost hole.
1124 * If yes, we restart sending from still existing holes,
1125 * and adjust cwnd via the congestion control module.
1126 */
1127 void
tcp_sack_lost_retransmission(struct tcpcb * tp,struct tcphdr * th)1128 tcp_sack_lost_retransmission(struct tcpcb *tp, struct tcphdr *th)
1129 {
1130 struct sackhole *temp;
1131
1132 if (IN_RECOVERY(tp->t_flags) &&
1133 SEQ_GT(tp->snd_fack, tp->snd_recover) &&
1134 ((temp = TAILQ_FIRST(&tp->snd_holes)) != NULL) &&
1135 SEQ_GEQ(temp->rxmit, temp->end) &&
1136 SEQ_GEQ(tp->snd_fack, temp->rxmit)) {
1137 TCPSTAT_INC(tcps_sack_lostrexmt);
1138 /*
1139 * Start retransmissions from the first hole, and
1140 * subsequently all other remaining holes, including
1141 * those, which had been sent completely before.
1142 */
1143 tp->sackhint.nexthole = temp;
1144 TAILQ_FOREACH(temp, &tp->snd_holes, scblink) {
1145 if (SEQ_GEQ(tp->snd_fack, temp->rxmit) &&
1146 SEQ_GEQ(temp->rxmit, temp->end))
1147 temp->rxmit = temp->start;
1148 }
1149 /*
1150 * Remember the old ssthresh, to deduct the beta factor used
1151 * by the CC module. Finally, set cwnd to ssthresh just
1152 * prior to invoking another cwnd reduction by the CC
1153 * module, to not shrink it excessively.
1154 */
1155 tp->snd_cwnd = tp->snd_ssthresh;
1156 /*
1157 * Formally exit recovery, and let the CC module adjust
1158 * ssthresh as intended.
1159 */
1160 EXIT_RECOVERY(tp->t_flags);
1161 cc_cong_signal(tp, th, CC_NDUPACK);
1162 /*
1163 * For PRR, adjust recover_fs as if this new reduction
1164 * initialized this variable.
1165 * cwnd will be adjusted by SACK or PRR processing
1166 * subsequently, only set it to a safe value here.
1167 */
1168 tp->snd_cwnd = tcp_maxseg(tp);
1169 tp->sackhint.recover_fs = (tp->snd_max - tp->snd_una) -
1170 tp->sackhint.recover_fs;
1171 }
1172 }
1173