1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
|
/*-
* Copyright (c) 2007, Myricom Inc.
* Copyright (c) 2008, Intel Corporation.
* Copyright (c) 2012 The FreeBSD Foundation
* Copyright (c) 2016 Mellanox Technologies.
* All rights reserved.
*
* Portions of this software were developed by Bjoern Zeeb
* under sponsorship from the FreeBSD Foundation.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_inet.h"
#include "opt_inet6.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/mbuf.h>
#include <sys/kernel.h>
#include <sys/socket.h>
#include <net/if.h>
#include <net/if_var.h>
#include <net/ethernet.h>
#include <net/vnet.h>
#include <netinet/in_systm.h>
#include <netinet/in.h>
#include <netinet/ip6.h>
#include <netinet/ip.h>
#include <netinet/ip_var.h>
#include <netinet/tcp.h>
#include <netinet/tcp_lro.h>
#include <netinet6/ip6_var.h>
#include <machine/in_cksum.h>
static MALLOC_DEFINE(M_LRO, "LRO", "LRO control structures");
#define TCP_LRO_UPDATE_CSUM 1
#ifndef TCP_LRO_UPDATE_CSUM
#define TCP_LRO_INVALID_CSUM 0x0000
#endif
int
tcp_lro_init(struct lro_ctrl *lc)
{
return (tcp_lro_init_args(lc, NULL, TCP_LRO_ENTRIES, 0));
}
int
tcp_lro_init_args(struct lro_ctrl *lc, struct ifnet *ifp,
unsigned lro_entries, unsigned lro_mbufs)
{
struct lro_entry *le;
size_t size;
unsigned i;
lc->lro_bad_csum = 0;
lc->lro_queued = 0;
lc->lro_flushed = 0;
lc->lro_cnt = 0;
lc->lro_mbuf_count = 0;
lc->lro_mbuf_max = lro_mbufs;
lc->lro_cnt = lro_entries;
lc->ifp = ifp;
SLIST_INIT(&lc->lro_free);
SLIST_INIT(&lc->lro_active);
/* compute size to allocate */
size = (lro_mbufs * sizeof(struct mbuf *)) +
(lro_entries * sizeof(*le));
lc->lro_mbuf_data = (struct mbuf **)
malloc(size, M_LRO, M_NOWAIT | M_ZERO);
/* check for out of memory */
if (lc->lro_mbuf_data == NULL) {
memset(lc, 0, sizeof(*lc));
return (ENOMEM);
}
/* compute offset for LRO entries */
le = (struct lro_entry *)
(lc->lro_mbuf_data + lro_mbufs);
/* setup linked list */
for (i = 0; i != lro_entries; i++)
SLIST_INSERT_HEAD(&lc->lro_free, le + i, next);
return (0);
}
void
tcp_lro_free(struct lro_ctrl *lc)
{
struct lro_entry *le;
unsigned x;
/* reset LRO free list */
SLIST_INIT(&lc->lro_free);
/* free active mbufs, if any */
while ((le = SLIST_FIRST(&lc->lro_active)) != NULL) {
SLIST_REMOVE_HEAD(&lc->lro_active, next);
m_freem(le->m_head);
}
/* free mbuf array, if any */
for (x = 0; x != lc->lro_mbuf_count; x++)
m_freem(lc->lro_mbuf_data[x]);
lc->lro_mbuf_count = 0;
/* free allocated memory, if any */
free(lc->lro_mbuf_data, M_LRO);
lc->lro_mbuf_data = NULL;
}
#ifdef TCP_LRO_UPDATE_CSUM
static uint16_t
tcp_lro_csum_th(struct tcphdr *th)
{
uint32_t ch;
uint16_t *p, l;
ch = th->th_sum = 0x0000;
l = th->th_off;
p = (uint16_t *)th;
while (l > 0) {
ch += *p;
p++;
ch += *p;
p++;
l--;
}
while (ch > 0xffff)
ch = (ch >> 16) + (ch & 0xffff);
return (ch & 0xffff);
}
static uint16_t
tcp_lro_rx_csum_fixup(struct lro_entry *le, void *l3hdr, struct tcphdr *th,
uint16_t tcp_data_len, uint16_t csum)
{
uint32_t c;
uint16_t cs;
c = csum;
/* Remove length from checksum. */
switch (le->eh_type) {
#ifdef INET6
case ETHERTYPE_IPV6:
{
struct ip6_hdr *ip6;
ip6 = (struct ip6_hdr *)l3hdr;
if (le->append_cnt == 0)
cs = ip6->ip6_plen;
else {
uint32_t cx;
cx = ntohs(ip6->ip6_plen);
cs = in6_cksum_pseudo(ip6, cx, ip6->ip6_nxt, 0);
}
break;
}
#endif
#ifdef INET
case ETHERTYPE_IP:
{
struct ip *ip4;
ip4 = (struct ip *)l3hdr;
if (le->append_cnt == 0)
cs = ip4->ip_len;
else {
cs = in_addword(ntohs(ip4->ip_len) - sizeof(*ip4),
IPPROTO_TCP);
cs = in_pseudo(ip4->ip_src.s_addr, ip4->ip_dst.s_addr,
htons(cs));
}
break;
}
#endif
default:
cs = 0; /* Keep compiler happy. */
}
cs = ~cs;
c += cs;
/* Remove TCP header csum. */
cs = ~tcp_lro_csum_th(th);
c += cs;
while (c > 0xffff)
c = (c >> 16) + (c & 0xffff);
return (c & 0xffff);
}
#endif
void
tcp_lro_flush_inactive(struct lro_ctrl *lc, const struct timeval *timeout)
{
struct lro_entry *le, *le_tmp;
struct timeval tv;
if (SLIST_EMPTY(&lc->lro_active))
return;
getmicrotime(&tv);
timevalsub(&tv, timeout);
SLIST_FOREACH_SAFE(le, &lc->lro_active, next, le_tmp) {
if (timevalcmp(&tv, &le->mtime, >=)) {
SLIST_REMOVE(&lc->lro_active, le, lro_entry, next);
tcp_lro_flush(lc, le);
}
}
}
void
tcp_lro_flush(struct lro_ctrl *lc, struct lro_entry *le)
{
if (le->append_cnt > 0) {
struct tcphdr *th;
uint16_t p_len;
p_len = htons(le->p_len);
switch (le->eh_type) {
#ifdef INET6
case ETHERTYPE_IPV6:
{
struct ip6_hdr *ip6;
ip6 = le->le_ip6;
ip6->ip6_plen = p_len;
th = (struct tcphdr *)(ip6 + 1);
le->m_head->m_pkthdr.csum_flags = CSUM_DATA_VALID |
CSUM_PSEUDO_HDR;
le->p_len += ETHER_HDR_LEN + sizeof(*ip6);
break;
}
#endif
#ifdef INET
case ETHERTYPE_IP:
{
struct ip *ip4;
#ifdef TCP_LRO_UPDATE_CSUM
uint32_t cl;
uint16_t c;
#endif
ip4 = le->le_ip4;
#ifdef TCP_LRO_UPDATE_CSUM
/* Fix IP header checksum for new length. */
c = ~ip4->ip_sum;
cl = c;
c = ~ip4->ip_len;
cl += c + p_len;
while (cl > 0xffff)
cl = (cl >> 16) + (cl & 0xffff);
c = cl;
ip4->ip_sum = ~c;
#else
ip4->ip_sum = TCP_LRO_INVALID_CSUM;
#endif
ip4->ip_len = p_len;
th = (struct tcphdr *)(ip4 + 1);
le->m_head->m_pkthdr.csum_flags = CSUM_DATA_VALID |
CSUM_PSEUDO_HDR | CSUM_IP_CHECKED | CSUM_IP_VALID;
le->p_len += ETHER_HDR_LEN;
break;
}
#endif
default:
th = NULL; /* Keep compiler happy. */
}
le->m_head->m_pkthdr.csum_data = 0xffff;
le->m_head->m_pkthdr.len = le->p_len;
/* Incorporate the latest ACK into the TCP header. */
th->th_ack = le->ack_seq;
th->th_win = le->window;
/* Incorporate latest timestamp into the TCP header. */
if (le->timestamp != 0) {
uint32_t *ts_ptr;
ts_ptr = (uint32_t *)(th + 1);
ts_ptr[1] = htonl(le->tsval);
ts_ptr[2] = le->tsecr;
}
#ifdef TCP_LRO_UPDATE_CSUM
/* Update the TCP header checksum. */
le->ulp_csum += p_len;
le->ulp_csum += tcp_lro_csum_th(th);
while (le->ulp_csum > 0xffff)
le->ulp_csum = (le->ulp_csum >> 16) +
(le->ulp_csum & 0xffff);
th->th_sum = (le->ulp_csum & 0xffff);
th->th_sum = ~th->th_sum;
#else
th->th_sum = TCP_LRO_INVALID_CSUM;
#endif
}
(*lc->ifp->if_input)(lc->ifp, le->m_head);
lc->lro_queued += le->append_cnt + 1;
lc->lro_flushed++;
bzero(le, sizeof(*le));
SLIST_INSERT_HEAD(&lc->lro_free, le, next);
}
static int
tcp_lro_mbuf_compare_header(const void *ppa, const void *ppb)
{
const struct mbuf *ma = *((const struct mbuf * const *)ppa);
const struct mbuf *mb = *((const struct mbuf * const *)ppb);
int ret;
ret = M_HASHTYPE_GET(ma) - M_HASHTYPE_GET(mb);
if (ret != 0)
goto done;
ret = ma->m_pkthdr.flowid - mb->m_pkthdr.flowid;
if (ret != 0)
goto done;
ret = TCP_LRO_SEQUENCE(ma) - TCP_LRO_SEQUENCE(mb);
done:
return (ret);
}
void
tcp_lro_flush_all(struct lro_ctrl *lc)
{
struct lro_entry *le;
uint32_t hashtype;
uint32_t flowid;
unsigned x;
/* check if no mbufs to flush */
if (__predict_false(lc->lro_mbuf_count == 0))
goto done;
/* sort all mbufs according to stream */
qsort(lc->lro_mbuf_data, lc->lro_mbuf_count, sizeof(struct mbuf *),
&tcp_lro_mbuf_compare_header);
/* input data into LRO engine, stream by stream */
flowid = 0;
hashtype = M_HASHTYPE_NONE;
for (x = 0; x != lc->lro_mbuf_count; x++) {
struct mbuf *mb;
mb = lc->lro_mbuf_data[x];
/* check for new stream */
if (mb->m_pkthdr.flowid != flowid ||
M_HASHTYPE_GET(mb) != hashtype) {
flowid = mb->m_pkthdr.flowid;
hashtype = M_HASHTYPE_GET(mb);
/* flush active streams */
while ((le = SLIST_FIRST(&lc->lro_active)) != NULL) {
SLIST_REMOVE_HEAD(&lc->lro_active, next);
tcp_lro_flush(lc, le);
}
}
#ifdef TCP_LRO_RESET_SEQUENCE
/* reset sequence number */
TCP_LRO_SEQUENCE(mb) = 0;
#endif
/* add packet to LRO engine */
if (tcp_lro_rx(lc, mb, 0) != 0) {
/* input packet to network layer */
(*lc->ifp->if_input)(lc->ifp, mb);
lc->lro_queued++;
lc->lro_flushed++;
}
}
done:
/* flush active streams */
while ((le = SLIST_FIRST(&lc->lro_active)) != NULL) {
SLIST_REMOVE_HEAD(&lc->lro_active, next);
tcp_lro_flush(lc, le);
}
lc->lro_mbuf_count = 0;
}
#ifdef INET6
static int
tcp_lro_rx_ipv6(struct lro_ctrl *lc, struct mbuf *m, struct ip6_hdr *ip6,
struct tcphdr **th)
{
/* XXX-BZ we should check the flow-label. */
/* XXX-BZ We do not yet support ext. hdrs. */
if (ip6->ip6_nxt != IPPROTO_TCP)
return (TCP_LRO_NOT_SUPPORTED);
/* Find the TCP header. */
*th = (struct tcphdr *)(ip6 + 1);
return (0);
}
#endif
#ifdef INET
static int
tcp_lro_rx_ipv4(struct lro_ctrl *lc, struct mbuf *m, struct ip *ip4,
struct tcphdr **th)
{
int csum_flags;
uint16_t csum;
if (ip4->ip_p != IPPROTO_TCP)
return (TCP_LRO_NOT_SUPPORTED);
/* Ensure there are no options. */
if ((ip4->ip_hl << 2) != sizeof (*ip4))
return (TCP_LRO_CANNOT);
/* .. and the packet is not fragmented. */
if (ip4->ip_off & htons(IP_MF|IP_OFFMASK))
return (TCP_LRO_CANNOT);
/* Legacy IP has a header checksum that needs to be correct. */
csum_flags = m->m_pkthdr.csum_flags;
if (csum_flags & CSUM_IP_CHECKED) {
if (__predict_false((csum_flags & CSUM_IP_VALID) == 0)) {
lc->lro_bad_csum++;
return (TCP_LRO_CANNOT);
}
} else {
csum = in_cksum_hdr(ip4);
if (__predict_false((csum) != 0)) {
lc->lro_bad_csum++;
return (TCP_LRO_CANNOT);
}
}
/* Find the TCP header (we assured there are no IP options). */
*th = (struct tcphdr *)(ip4 + 1);
return (0);
}
#endif
int
tcp_lro_rx(struct lro_ctrl *lc, struct mbuf *m, uint32_t csum)
{
struct lro_entry *le;
struct ether_header *eh;
#ifdef INET6
struct ip6_hdr *ip6 = NULL; /* Keep compiler happy. */
#endif
#ifdef INET
struct ip *ip4 = NULL; /* Keep compiler happy. */
#endif
struct tcphdr *th;
void *l3hdr = NULL; /* Keep compiler happy. */
uint32_t *ts_ptr;
tcp_seq seq;
int error, ip_len, l;
uint16_t eh_type, tcp_data_len;
/* We expect a contiguous header [eh, ip, tcp]. */
eh = mtod(m, struct ether_header *);
eh_type = ntohs(eh->ether_type);
switch (eh_type) {
#ifdef INET6
case ETHERTYPE_IPV6:
{
CURVNET_SET(lc->ifp->if_vnet);
if (V_ip6_forwarding != 0) {
/* XXX-BZ stats but changing lro_ctrl is a problem. */
CURVNET_RESTORE();
return (TCP_LRO_CANNOT);
}
CURVNET_RESTORE();
l3hdr = ip6 = (struct ip6_hdr *)(eh + 1);
error = tcp_lro_rx_ipv6(lc, m, ip6, &th);
if (error != 0)
return (error);
tcp_data_len = ntohs(ip6->ip6_plen);
ip_len = sizeof(*ip6) + tcp_data_len;
break;
}
#endif
#ifdef INET
case ETHERTYPE_IP:
{
CURVNET_SET(lc->ifp->if_vnet);
if (V_ipforwarding != 0) {
/* XXX-BZ stats but changing lro_ctrl is a problem. */
CURVNET_RESTORE();
return (TCP_LRO_CANNOT);
}
CURVNET_RESTORE();
l3hdr = ip4 = (struct ip *)(eh + 1);
error = tcp_lro_rx_ipv4(lc, m, ip4, &th);
if (error != 0)
return (error);
ip_len = ntohs(ip4->ip_len);
tcp_data_len = ip_len - sizeof(*ip4);
break;
}
#endif
/* XXX-BZ what happens in case of VLAN(s)? */
default:
return (TCP_LRO_NOT_SUPPORTED);
}
/*
* If the frame is padded beyond the end of the IP packet, then we must
* trim the extra bytes off.
*/
l = m->m_pkthdr.len - (ETHER_HDR_LEN + ip_len);
if (l != 0) {
if (l < 0)
/* Truncated packet. */
return (TCP_LRO_CANNOT);
m_adj(m, -l);
}
/*
* Check TCP header constraints.
*/
/* Ensure no bits set besides ACK or PSH. */
if ((th->th_flags & ~(TH_ACK | TH_PUSH)) != 0)
return (TCP_LRO_CANNOT);
/* XXX-BZ We lose a AKC|PUSH flag concatinating multiple segments. */
/* XXX-BZ Ideally we'd flush on PUSH? */
/*
* Check for timestamps.
* Since the only option we handle are timestamps, we only have to
* handle the simple case of aligned timestamps.
*/
l = (th->th_off << 2);
tcp_data_len -= l;
l -= sizeof(*th);
ts_ptr = (uint32_t *)(th + 1);
if (l != 0 && (__predict_false(l != TCPOLEN_TSTAMP_APPA) ||
(*ts_ptr != ntohl(TCPOPT_NOP<<24|TCPOPT_NOP<<16|
TCPOPT_TIMESTAMP<<8|TCPOLEN_TIMESTAMP))))
return (TCP_LRO_CANNOT);
/* If the driver did not pass in the checksum, set it now. */
if (csum == 0x0000)
csum = th->th_sum;
seq = ntohl(th->th_seq);
/* Try to find a matching previous segment. */
SLIST_FOREACH(le, &lc->lro_active, next) {
if (le->eh_type != eh_type)
continue;
if (le->source_port != th->th_sport ||
le->dest_port != th->th_dport)
continue;
switch (eh_type) {
#ifdef INET6
case ETHERTYPE_IPV6:
if (bcmp(&le->source_ip6, &ip6->ip6_src,
sizeof(struct in6_addr)) != 0 ||
bcmp(&le->dest_ip6, &ip6->ip6_dst,
sizeof(struct in6_addr)) != 0)
continue;
break;
#endif
#ifdef INET
case ETHERTYPE_IP:
if (le->source_ip4 != ip4->ip_src.s_addr ||
le->dest_ip4 != ip4->ip_dst.s_addr)
continue;
break;
#endif
}
/* Flush now if appending will result in overflow. */
if (le->p_len > (65535 - tcp_data_len)) {
SLIST_REMOVE(&lc->lro_active, le, lro_entry, next);
tcp_lro_flush(lc, le);
break;
}
/* Try to append the new segment. */
if (__predict_false(seq != le->next_seq ||
(tcp_data_len == 0 && le->ack_seq == th->th_ack))) {
/* Out of order packet or duplicate ACK. */
SLIST_REMOVE(&lc->lro_active, le, lro_entry, next);
tcp_lro_flush(lc, le);
return (TCP_LRO_CANNOT);
}
if (l != 0) {
uint32_t tsval = ntohl(*(ts_ptr + 1));
/* Make sure timestamp values are increasing. */
/* XXX-BZ flip and use TSTMP_GEQ macro for this? */
if (__predict_false(le->tsval > tsval ||
*(ts_ptr + 2) == 0))
return (TCP_LRO_CANNOT);
le->tsval = tsval;
le->tsecr = *(ts_ptr + 2);
}
le->next_seq += tcp_data_len;
le->ack_seq = th->th_ack;
le->window = th->th_win;
le->append_cnt++;
#ifdef TCP_LRO_UPDATE_CSUM
le->ulp_csum += tcp_lro_rx_csum_fixup(le, l3hdr, th,
tcp_data_len, ~csum);
#endif
if (tcp_data_len == 0) {
m_freem(m);
return (0);
}
le->p_len += tcp_data_len;
/*
* Adjust the mbuf so that m_data points to the first byte of
* the ULP payload. Adjust the mbuf to avoid complications and
* append new segment to existing mbuf chain.
*/
m_adj(m, m->m_pkthdr.len - tcp_data_len);
m_demote_pkthdr(m);
le->m_tail->m_next = m;
le->m_tail = m_last(m);
/*
* If a possible next full length packet would cause an
* overflow, pro-actively flush now.
*/
if (le->p_len > (65535 - lc->ifp->if_mtu)) {
SLIST_REMOVE(&lc->lro_active, le, lro_entry, next);
tcp_lro_flush(lc, le);
} else
getmicrotime(&le->mtime);
return (0);
}
/* Try to find an empty slot. */
if (SLIST_EMPTY(&lc->lro_free))
return (TCP_LRO_CANNOT);
/* Start a new segment chain. */
le = SLIST_FIRST(&lc->lro_free);
SLIST_REMOVE_HEAD(&lc->lro_free, next);
SLIST_INSERT_HEAD(&lc->lro_active, le, next);
getmicrotime(&le->mtime);
/* Start filling in details. */
switch (eh_type) {
#ifdef INET6
case ETHERTYPE_IPV6:
le->le_ip6 = ip6;
le->source_ip6 = ip6->ip6_src;
le->dest_ip6 = ip6->ip6_dst;
le->eh_type = eh_type;
le->p_len = m->m_pkthdr.len - ETHER_HDR_LEN - sizeof(*ip6);
break;
#endif
#ifdef INET
case ETHERTYPE_IP:
le->le_ip4 = ip4;
le->source_ip4 = ip4->ip_src.s_addr;
le->dest_ip4 = ip4->ip_dst.s_addr;
le->eh_type = eh_type;
le->p_len = m->m_pkthdr.len - ETHER_HDR_LEN;
break;
#endif
}
le->source_port = th->th_sport;
le->dest_port = th->th_dport;
le->next_seq = seq + tcp_data_len;
le->ack_seq = th->th_ack;
le->window = th->th_win;
if (l != 0) {
le->timestamp = 1;
le->tsval = ntohl(*(ts_ptr + 1));
le->tsecr = *(ts_ptr + 2);
}
#ifdef TCP_LRO_UPDATE_CSUM
/*
* Do not touch the csum of the first packet. However save the
* "adjusted" checksum of just the source and destination addresses,
* the next header and the TCP payload. The length and TCP header
* parts may change, so we remove those from the saved checksum and
* re-add with final values on tcp_lro_flush() if needed.
*/
KASSERT(le->ulp_csum == 0, ("%s: le=%p le->ulp_csum=0x%04x\n",
__func__, le, le->ulp_csum));
le->ulp_csum = tcp_lro_rx_csum_fixup(le, l3hdr, th, tcp_data_len,
~csum);
th->th_sum = csum; /* Restore checksum on first packet. */
#endif
le->m_head = m;
le->m_tail = m_last(m);
return (0);
}
void
tcp_lro_queue_mbuf(struct lro_ctrl *lc, struct mbuf *mb)
{
/* sanity checks */
if (__predict_false(lc->ifp == NULL || lc->lro_mbuf_data == NULL ||
lc->lro_mbuf_max == 0)) {
/* packet drop */
m_freem(mb);
return;
}
/* check if packet is not LRO capable */
if (__predict_false(mb->m_pkthdr.csum_flags == 0 ||
(lc->ifp->if_capenable & IFCAP_LRO) == 0)) {
lc->lro_flushed++;
lc->lro_queued++;
/* input packet to network layer */
(*lc->ifp->if_input) (lc->ifp, mb);
return;
}
/* check if array is full */
if (__predict_false(lc->lro_mbuf_count == lc->lro_mbuf_max))
tcp_lro_flush_all(lc);
/* store sequence number */
TCP_LRO_SEQUENCE(mb) = lc->lro_mbuf_count;
/* enter mbuf */
lc->lro_mbuf_data[lc->lro_mbuf_count++] = mb;
}
/* end */
|