aboutgitcodebugslistschat
path: root/udp.c
blob: 4d612c3180a749b32aa2c2538a4ff976d7f5785b (plain) (blame)
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
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
// SPDX-License-Identifier: GPL-2.0-or-later

/* PASST - Plug A Simple Socket Transport
 *  for qemu/UNIX domain socket mode
 *
 * PASTA - Pack A Subtle Tap Abstraction
 *  for network namespace/tap device mode
 *
 * udp.c - UDP L2-L4 translation routines
 *
 * Copyright (c) 2020-2021 Red Hat GmbH
 * Author: Stefano Brivio <sbrivio@redhat.com>
 */

/**
 * DOC: Theory of Operation
 *
 * UDP Flows
 * =========
 *
 * UDP doesn't have true connections, but many protocols use a connection-like
 * format.  The flow is initiated by a client sending a datagram from a port of
 * its choosing (usually ephemeral) to a specific port (usually well known) on a
 * server.  Both client and server address must be unicast.  The server sends
 * replies using the same addresses & ports with src/dest swapped.
 *
 * We track pseudo-connections of this type as flow table entries of type
 * FLOW_UDP.  We store the time of the last traffic on the flow in uflow->ts,
 * and let the flow expire if there is no traffic for UDP_CONN_TIMEOUT seconds.
 *
 * NOTE: This won't handle multicast protocols, or some protocols with different
 * port usage.  We'll need specific logic if we want to handle those.
 *
 * "Listening" sockets
 * ===================
 *
 * UDP doesn't use listen(), but we consider long term sockets which are allowed
 * to create new flows "listening" by analogy with TCP. This listening socket
 * could receive packets from multiple flows, so we use a hash table match to
 * find the specific flow for a datagram.
 *
 * When a UDP flow is initiated from a listening socket we take a duplicate of
 * the socket and store it in uflow->s[INISIDE].  This will last for the
 * lifetime of the flow, even if the original listening socket is closed due to
 * port auto-probing.  The duplicate is used to deliver replies back to the
 * originating side.
 *
 * Reply sockets
 * =============
 *
 * When a UDP flow targets a socket, we create a "reply" socket in
 * uflow->s[TGTSIDE] both to deliver datagrams to the target side and receive
 * replies on the target side.  This socket is both bound and connected and has
 * EPOLL_TYPE_UDP_REPLY.  The connect() means it will only receive datagrams
 * associated with this flow, so the epoll reference directly points to the flow
 * and we don't need a hash lookup.
 *
 * NOTE: it's possible that the reply socket could have a bound address
 * overlapping with an unrelated listening socket.  We assume datagrams for the
 * flow will come to the reply socket in preference to a listening socket.  The
 * sample program doc/platform-requirements/reuseaddr-priority.c documents and
 * tests that assumption.
 *
 * "Spliced" flows
 * ===============
 *
 * In PASTA mode, L2-L4 translation is skipped for connections to ports bound
 * between namespaces using the loopback interface, messages are directly
 * transferred between L4 sockets instead. These are called spliced connections
 * in analogy with the TCP implementation.  The the splice() syscall isn't
 * actually used; it doesn't make sense for datagrams and instead a pair of
 * recvmmsg() and sendmmsg() is used to forward the datagrams.
 *
 * Note that a spliced flow will have *both* a duplicated listening socket and a
 * reply socket (see above).
 */

#include <sched.h>
#include <unistd.h>
#include <signal.h>
#include <stdio.h>
#include <errno.h>
#include <limits.h>
#include <assert.h>
#include <net/ethernet.h>
#include <net/if.h>
#include <netinet/in.h>
#include <netinet/ip.h>
#include <netinet/udp.h>
#include <stdint.h>
#include <stddef.h>
#include <string.h>
#include <sys/epoll.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <sys/uio.h>
#include <time.h>
#include <fcntl.h>
#include <arpa/inet.h>
#include <linux/errqueue.h>

#include "checksum.h"
#include "util.h"
#include "iov.h"
#include "ip.h"
#include "siphash.h"
#include "inany.h"
#include "passt.h"
#include "tap.h"
#include "pcap.h"
#include "log.h"
#include "flow_table.h"

#define UDP_CONN_TIMEOUT	180 /* s, timeout for ephemeral or local bind */
#define UDP_MAX_FRAMES		32  /* max # of frames to receive at once */

/* "Spliced" sockets indexed by bound port (host order) */
static int udp_splice_ns  [IP_VERSIONS][NUM_PORTS];
static int udp_splice_init[IP_VERSIONS][NUM_PORTS];

/* Static buffers */

/**
 * struct udp_payload_t - UDP header and data for inbound messages
 * @uh:		UDP header
 * @data:	UDP data
 */
static struct udp_payload_t {
	struct udphdr uh;
	char data[USHRT_MAX - sizeof(struct udphdr)];
#ifdef __AVX2__
} __attribute__ ((packed, aligned(32)))
#else
} __attribute__ ((packed, aligned(__alignof__(unsigned int))))
#endif
udp_payload[UDP_MAX_FRAMES];

/* Ethernet header for IPv4 frames */
static struct ethhdr udp4_eth_hdr;

/* Ethernet header for IPv6 frames */
static struct ethhdr udp6_eth_hdr;

/**
 * struct udp_meta_t - Pre-cooked headers and metadata for UDP packets
 * @ip6h:	Pre-filled IPv6 header (except for payload_len and addresses)
 * @ip4h:	Pre-filled IPv4 header (except for tot_len and saddr)
 * @taph:	Tap backend specific header
 * @s_in:	Source socket address, filled in by recvmmsg()
 * @tosidx:	sidx for the destination side of this datagram's flow
 */
static struct udp_meta_t {
	struct ipv6hdr ip6h;
	struct iphdr ip4h;
	struct tap_hdr taph;

	union sockaddr_inany s_in;
	flow_sidx_t tosidx;
}
#ifdef __AVX2__
__attribute__ ((aligned(32)))
#endif
udp_meta[UDP_MAX_FRAMES];

/**
 * enum udp_iov_idx - Indices for the buffers making up a single UDP frame
 * @UDP_IOV_TAP         tap specific header
 * @UDP_IOV_ETH         Ethernet header
 * @UDP_IOV_IP          IP (v4/v6) header
 * @UDP_IOV_PAYLOAD     IP payload (UDP header + data)
 * @UDP_NUM_IOVS        the number of entries in the iovec array
 */
enum udp_iov_idx {
	UDP_IOV_TAP	= 0,
	UDP_IOV_ETH	= 1,
	UDP_IOV_IP	= 2,
	UDP_IOV_PAYLOAD	= 3,
	UDP_NUM_IOVS
};

/* IOVs and msghdr arrays for receiving datagrams from sockets */
static struct iovec	udp_iov_recv		[UDP_MAX_FRAMES];
static struct mmsghdr	udp4_mh_recv		[UDP_MAX_FRAMES];
static struct mmsghdr	udp6_mh_recv		[UDP_MAX_FRAMES];

/* IOVs and msghdr arrays for sending "spliced" datagrams to sockets */
static union sockaddr_inany udp_splice_to;

static struct iovec	udp_iov_splice		[UDP_MAX_FRAMES];
static struct mmsghdr	udp_mh_splice		[UDP_MAX_FRAMES];

/* IOVs for L2 frames */
static struct iovec	udp_l2_iov		[UDP_MAX_FRAMES][UDP_NUM_IOVS];

/**
 * udp_portmap_clear() - Clear UDP port map before configuration
 */
void udp_portmap_clear(void)
{
	unsigned i;

	for (i = 0; i < NUM_PORTS; i++) {
		udp_splice_ns[V4][i] = udp_splice_ns[V6][i] = -1;
		udp_splice_init[V4][i] = udp_splice_init[V6][i] = -1;
	}
}

/**
 * udp_update_l2_buf() - Update L2 buffers with Ethernet and IPv4 addresses
 * @eth_d:	Ethernet destination address, NULL if unchanged
 * @eth_s:	Ethernet source address, NULL if unchanged
 */
void udp_update_l2_buf(const unsigned char *eth_d, const unsigned char *eth_s)
{
	eth_update_mac(&udp4_eth_hdr, eth_d, eth_s);
	eth_update_mac(&udp6_eth_hdr, eth_d, eth_s);
}

/**
 * udp_iov_init_one() - Initialise scatter-gather lists for one buffer
 * @c:		Execution context
 * @i:		Index of buffer to initialize
 */
static void udp_iov_init_one(const struct ctx *c, size_t i)
{
	struct udp_payload_t *payload = &udp_payload[i];
	struct udp_meta_t *meta = &udp_meta[i];
	struct iovec *siov = &udp_iov_recv[i];
	struct iovec *tiov = udp_l2_iov[i];

	*meta = (struct udp_meta_t) {
		.ip4h = L2_BUF_IP4_INIT(IPPROTO_UDP),
		.ip6h = L2_BUF_IP6_INIT(IPPROTO_UDP),
	};

	*siov = IOV_OF_LVALUE(payload->data);

	tiov[UDP_IOV_TAP] = tap_hdr_iov(c, &meta->taph);
	tiov[UDP_IOV_PAYLOAD].iov_base = payload;

	/* It's useful to have separate msghdr arrays for receiving.  Otherwise,
	 * an IPv4 recv() will alter msg_namelen, so we'd have to reset it every
	 * time or risk truncating the address on future IPv6 recv()s.
	 */
	if (c->ifi4) {
		struct msghdr *mh = &udp4_mh_recv[i].msg_hdr;

		mh->msg_name	= &meta->s_in;
		mh->msg_namelen	= sizeof(struct sockaddr_in);
		mh->msg_iov	= siov;
		mh->msg_iovlen	= 1;
	}

	if (c->ifi6) {
		struct msghdr *mh = &udp6_mh_recv[i].msg_hdr;

		mh->msg_name	= &meta->s_in;
		mh->msg_namelen	= sizeof(struct sockaddr_in6);
		mh->msg_iov	= siov;
		mh->msg_iovlen	= 1;
	}
}

/**
 * udp_iov_init() - Initialise scatter-gather L2 buffers
 * @c:		Execution context
 */
static void udp_iov_init(const struct ctx *c)
{
	size_t i;

	udp4_eth_hdr.h_proto = htons_constant(ETH_P_IP);
	udp6_eth_hdr.h_proto = htons_constant(ETH_P_IPV6);

	for (i = 0; i < UDP_MAX_FRAMES; i++)
		udp_iov_init_one(c, i);
}

/**
 * udp_at_sidx() - Get UDP specific flow at given sidx
 * @sidx:    Flow and side to retrieve
 *
 * Return: UDP specific flow at @sidx, or NULL of @sidx is invalid.  Asserts if
 *         the flow at @sidx is not FLOW_UDP.
 */
struct udp_flow *udp_at_sidx(flow_sidx_t sidx)
{
	union flow *flow = flow_at_sidx(sidx);

	if (!flow)
		return NULL;

	ASSERT(flow->f.type == FLOW_UDP);
	return &flow->udp;
}

/*
 * udp_flow_close() - Close and clean up UDP flow
 * @c:		Execution context
 * @uflow:	UDP flow
 */
static void udp_flow_close(const struct ctx *c, struct udp_flow *uflow)
{
	if (uflow->s[INISIDE] >= 0) {
		/* The listening socket needs to stay in epoll */
		close(uflow->s[INISIDE]);
		uflow->s[INISIDE] = -1;
	}

	if (uflow->s[TGTSIDE] >= 0) {
		/* But the flow specific one needs to be removed */
		epoll_ctl(c->epollfd, EPOLL_CTL_DEL, uflow->s[TGTSIDE], NULL);
		close(uflow->s[TGTSIDE]);
		uflow->s[TGTSIDE] = -1;
	}
	flow_hash_remove(c, FLOW_SIDX(uflow, INISIDE));
	if (!pif_is_socket(uflow->f.pif[TGTSIDE]))
		flow_hash_remove(c, FLOW_SIDX(uflow, TGTSIDE));
}

/**
 * udp_flow_new() - Common setup for a new UDP flow
 * @c:		Execution context
 * @flow:	Initiated flow
 * @s_ini:	Initiating socket (or -1)
 * @now:	Timestamp
 *
 * Return: UDP specific flow, if successful, NULL on failure
 */
static flow_sidx_t udp_flow_new(const struct ctx *c, union flow *flow,
				int s_ini, const struct timespec *now)
{
	const struct flowside *ini = &flow->f.side[INISIDE];
	struct udp_flow *uflow = NULL;
	const struct flowside *tgt;
	uint8_t tgtpif;

	if (!inany_is_unicast(&ini->eaddr) || ini->eport == 0) {
		flow_trace(flow, "Invalid endpoint to initiate UDP flow");
		goto cancel;
	}

	if (!(tgt = flow_target(c, flow, IPPROTO_UDP)))
		goto cancel;
	tgtpif = flow->f.pif[TGTSIDE];

	uflow = FLOW_SET_TYPE(flow, FLOW_UDP, udp);
	uflow->ts = now->tv_sec;
	uflow->s[INISIDE] = uflow->s[TGTSIDE] = -1;

	if (s_ini >= 0) {
		/* When using auto port-scanning the listening port could go
		 * away, so we need to duplicate the socket
		 */
		uflow->s[INISIDE] = fcntl(s_ini, F_DUPFD_CLOEXEC, 0);
		if (uflow->s[INISIDE] < 0) {
			flow_err(uflow,
				 "Couldn't duplicate listening socket: %s",
				 strerror(errno));
			goto cancel;
		}
	}

	if (pif_is_socket(tgtpif)) {
		struct mmsghdr discard[UIO_MAXIOV] = { 0 };
		union {
			flow_sidx_t sidx;
			uint32_t data;
		} fref = {
			.sidx = FLOW_SIDX(flow, TGTSIDE),
		};
		int rc;

		uflow->s[TGTSIDE] = flowside_sock_l4(c, EPOLL_TYPE_UDP_REPLY,
						     tgtpif, tgt, fref.data);
		if (uflow->s[TGTSIDE] < 0) {
			flow_dbg(uflow,
				 "Couldn't open socket for spliced flow: %s",
				 strerror(errno));
			goto cancel;
		}

		if (flowside_connect(c, uflow->s[TGTSIDE], tgtpif, tgt) < 0) {
			flow_dbg(uflow,
				 "Couldn't connect flow socket: %s",
				 strerror(errno));
			goto cancel;
		}

		/* It's possible, if unlikely, that we could receive some
		 * unrelated packets in between the bind() and connect() of this
		 * socket.  For now we just discard these.  We could consider
		 * trying to redirect these to an appropriate handler, if we
		 * need to.
		 */
		rc = recvmmsg(uflow->s[TGTSIDE], discard, ARRAY_SIZE(discard),
			      MSG_DONTWAIT, NULL);
		if (rc >= ARRAY_SIZE(discard)) {
			flow_dbg(uflow,
				 "Too many (%d) spurious reply datagrams", rc);
			goto cancel;
		} else if (rc > 0) {
			flow_trace(uflow,
				   "Discarded %d spurious reply datagrams", rc);
		} else if (errno != EAGAIN) {
			flow_err(uflow,
				 "Unexpected error discarding datagrams: %s",
				 strerror(errno));
		}
	}

	flow_hash_insert(c, FLOW_SIDX(uflow, INISIDE));

	/* If the target side is a socket, it will be a reply socket that knows
	 * its own flowside.  But if it's tap, then we need to look it up by
	 * hash.
	 */
	if (!pif_is_socket(tgtpif))
		flow_hash_insert(c, FLOW_SIDX(uflow, TGTSIDE));
	FLOW_ACTIVATE(uflow);

	return FLOW_SIDX(uflow, TGTSIDE);

cancel:
	if (uflow)
		udp_flow_close(c, uflow);
	flow_alloc_cancel(flow);
	return FLOW_SIDX_NONE;
}

/**
 * udp_flow_from_sock() - Find or create UDP flow for "listening" socket
 * @c:		Execution context
 * @ref:	epoll reference of the receiving socket
 * @meta:	Metadata buffer for the datagram
 * @now:	Timestamp
 *
 * #syscalls fcntl
 *
 * Return: sidx for the destination side of the flow for this packet, or
 *         FLOW_SIDX_NONE if we couldn't find or create a flow.
 */
static flow_sidx_t udp_flow_from_sock(const struct ctx *c, union epoll_ref ref,
				      struct udp_meta_t *meta,
				      const struct timespec *now)
{
	struct udp_flow *uflow;
	union flow *flow;
	flow_sidx_t sidx;

	ASSERT(ref.type == EPOLL_TYPE_UDP);

	if (!ref.udp.orig)
		return FLOW_SIDX_NONE;

	sidx = flow_lookup_sa(c, IPPROTO_UDP, ref.udp.pif, &meta->s_in, ref.udp.port);
	if ((uflow = udp_at_sidx(sidx))) {
		uflow->ts = now->tv_sec;
		return flow_sidx_opposite(sidx);
	}

	if (!(flow = flow_alloc())) {
		char sastr[SOCKADDR_STRLEN];

		debug("Couldn't allocate flow for UDP datagram from %s %s",
		      pif_name(ref.udp.pif),
		      sockaddr_ntop(&meta->s_in, sastr, sizeof(sastr)));
		return FLOW_SIDX_NONE;
	}

	flow_initiate_sa(flow, ref.udp.pif, &meta->s_in, ref.udp.port);
	return udp_flow_new(c, flow, ref.fd, now);
}

/**
 * udp_splice_prepare() - Prepare one datagram for splicing
 * @mmh:	Receiving mmsghdr array
 * @idx:	Index of the datagram to prepare
 */
static void udp_splice_prepare(struct mmsghdr *mmh, unsigned idx)
{
	udp_mh_splice[idx].msg_hdr.msg_iov->iov_len = mmh[idx].msg_len;
}

/**
 * udp_splice_send() - Send a batch of datagrams from socket to socket
 * @c:		Execution context
 * @start:	Index of batch's first datagram in udp[46]_l2_buf
 * @n:		Number of datagrams in batch
 * @src:	Source port for datagram (target side)
 * @dst:	Destination port for datagrams (target side)
 * @ref:	epoll reference for origin socket
 * @now:	Timestamp
 */
static void udp_splice_send(const struct ctx *c, size_t start, size_t n,
			    flow_sidx_t tosidx)
{
	const struct flowside *toside = flowside_at_sidx(tosidx);
	const struct udp_flow *uflow = udp_at_sidx(tosidx);
	uint8_t topif = pif_at_sidx(tosidx);
	int s = uflow->s[tosidx.sidei];
	socklen_t sl;

	pif_sockaddr(c, &udp_splice_to, &sl, topif,
		     &toside->eaddr, toside->eport);

	sendmmsg(s, udp_mh_splice + start, n, MSG_NOSIGNAL);
}

/**
 * udp_update_hdr4() - Update headers for one IPv4 datagram
 * @ip4h:	Pre-filled IPv4 header (except for tot_len and saddr)
 * @bp:		Pointer to udp_payload_t to update
 * @toside:	Flowside for destination side
 * @dlen:	Length of UDP payload
 *
 * Return: size of IPv4 payload (UDP header + data)
 */
static size_t udp_update_hdr4(struct iphdr *ip4h, struct udp_payload_t *bp,
			      const struct flowside *toside, size_t dlen)
{
	const struct in_addr *src = inany_v4(&toside->faddr);
	const struct in_addr *dst = inany_v4(&toside->eaddr);
	size_t l4len = dlen + sizeof(bp->uh);
	size_t l3len = l4len + sizeof(*ip4h);

	ASSERT(src && dst);

	ip4h->tot_len = htons(l3len);
	ip4h->daddr = dst->s_addr;
	ip4h->saddr = src->s_addr;
	ip4h->check = csum_ip4_header(l3len, IPPROTO_UDP, *src, *dst);

	bp->uh.source = htons(toside->fport);
	bp->uh.dest = htons(toside->eport);
	bp->uh.len = htons(l4len);
	csum_udp4(&bp->uh, *src, *dst, bp->data, dlen);

	return l4len;
}

/**
 * udp_update_hdr6() - Update headers for one IPv6 datagram
 * @ip6h:	Pre-filled IPv6 header (except for payload_len and addresses)
 * @bp:		Pointer to udp_payload_t to update
 * @toside:	Flowside for destination side
 * @dlen:	Length of UDP payload
 *
 * Return: size of IPv6 payload (UDP header + data)
 */
static size_t udp_update_hdr6(struct ipv6hdr *ip6h, struct udp_payload_t *bp,
			      const struct flowside *toside, size_t dlen)
{
	uint16_t l4len = dlen + sizeof(bp->uh);

	ip6h->payload_len = htons(l4len);
	ip6h->daddr = toside->eaddr.a6;
	ip6h->saddr = toside->faddr.a6;
	ip6h->version = 6;
	ip6h->nexthdr = IPPROTO_UDP;
	ip6h->hop_limit = 255;

	bp->uh.source = htons(toside->fport);
	bp->uh.dest = htons(toside->eport);
	bp->uh.len = ip6h->payload_len;
	csum_udp6(&bp->uh, &toside->faddr.a6, &toside->eaddr.a6, bp->data, dlen);

	return l4len;
}

/**
 * udp_tap_prepare() - Convert one datagram into a tap frame
 * @mmh:	Receiving mmsghdr array
 * @idx:	Index of the datagram to prepare
 * @toside:	Flowside for destination side
 */
static void udp_tap_prepare(const struct mmsghdr *mmh, unsigned idx,
			    const struct flowside *toside)
{
	struct iovec (*tap_iov)[UDP_NUM_IOVS] = &udp_l2_iov[idx];
	struct udp_payload_t *bp = &udp_payload[idx];
	struct udp_meta_t *bm = &udp_meta[idx];
	size_t l4len;

	if (!inany_v4(&toside->eaddr) || !inany_v4(&toside->faddr)) {
		l4len = udp_update_hdr6(&bm->ip6h, bp, toside, mmh[idx].msg_len);
		tap_hdr_update(&bm->taph, l4len + sizeof(bm->ip6h) +
			       sizeof(udp6_eth_hdr));
		(*tap_iov)[UDP_IOV_ETH] = IOV_OF_LVALUE(udp6_eth_hdr);
		(*tap_iov)[UDP_IOV_IP] = IOV_OF_LVALUE(bm->ip6h);
	} else {
		l4len = udp_update_hdr4(&bm->ip4h, bp, toside, mmh[idx].msg_len);
		tap_hdr_update(&bm->taph, l4len + sizeof(bm->ip4h) +
			       sizeof(udp4_eth_hdr));
		(*tap_iov)[UDP_IOV_ETH] = IOV_OF_LVALUE(udp4_eth_hdr);
		(*tap_iov)[UDP_IOV_IP] = IOV_OF_LVALUE(bm->ip4h);
	}
	(*tap_iov)[UDP_IOV_PAYLOAD].iov_len = l4len;
}

/**
 * udp_sock_recverr() - Receive and clear an error from a socket
 * @s:		Socket to receive from
 *
 * Return: true if errors received and processed, false if no more errors
 *
 * #syscalls recvmsg
 */
static bool udp_sock_recverr(int s)
{
	const struct sock_extended_err *ee;
	const struct cmsghdr *hdr;
	char buf[CMSG_SPACE(sizeof(*ee))];
	struct msghdr mh = {
		.msg_name = NULL,
		.msg_namelen = 0,
		.msg_iov = NULL,
		.msg_iovlen = 0,
		.msg_control = buf,
		.msg_controllen = sizeof(buf),
	};
	ssize_t rc;

	rc = recvmsg(s, &mh, MSG_ERRQUEUE);
	if (rc < 0) {
		if (errno != EAGAIN && errno != EWOULDBLOCK)
			err_perror("Failed to read error queue");
		return false;
	}

	if (!(mh.msg_flags & MSG_ERRQUEUE)) {
		err("Missing MSG_ERRQUEUE flag reading error queue");
		return false;
	}

	hdr = CMSG_FIRSTHDR(&mh);
	if (!((hdr->cmsg_level == IPPROTO_IP &&
	       hdr->cmsg_type == IP_RECVERR) ||
	      (hdr->cmsg_level == IPPROTO_IPV6 &&
	       hdr->cmsg_type == IPV6_RECVERR))) {
		err("Unexpected cmsg reading error queue");
		return false;
	}

	ee = (const struct sock_extended_err *)CMSG_DATA(hdr);

	/* TODO: When possible propagate and otherwise handle errors */
	debug("%s error on UDP socket %i: %s",
	      str_ee_origin(ee), s, strerror(ee->ee_errno));

	return true;
}

/**
 * udp_sock_recv() - Receive datagrams from a socket
 * @c:		Execution context
 * @s:		Socket to receive from
 * @events:	epoll events bitmap
 * @mmh		mmsghdr array to receive into
 *
 * #syscalls recvmmsg
 */
static int udp_sock_recv(const struct ctx *c, int s, uint32_t events,
			 struct mmsghdr *mmh)
{
	/* For not entirely clear reasons (data locality?) pasta gets better
	 * throughput if we receive tap datagrams one at a atime.  For small
	 * splice datagrams throughput is slightly better if we do batch, but
	 * it's slightly worse for large splice datagrams.  Since we don't know
	 * before we receive whether we'll use tap or splice, always go one at a
	 * time for pasta mode.
	 */
	int n = (c->mode == MODE_PASTA ? 1 : UDP_MAX_FRAMES);

	ASSERT(!c->no_udp);

	/* Clear any errors first */
	if (events & EPOLLERR) {
		while (udp_sock_recverr(s))
			;
	}

	if (!(events & EPOLLIN))
		return 0;

	n = recvmmsg(s, mmh, n, 0, NULL);
	if (n < 0) {
		err_perror("Error receiving datagrams");
		return 0;
	}

	return n;
}

/**
 * udp_buf_sock_handler() - Handle new data from socket
 * @c:		Execution context
 * @ref:	epoll reference
 * @events:	epoll events bitmap
 * @now:	Current timestamp
 *
 * #syscalls recvmmsg
 */
void udp_buf_sock_handler(const struct ctx *c, union epoll_ref ref, uint32_t events,
			  const struct timespec *now)
{
	struct mmsghdr *mmh_recv = ref.udp.v6 ? udp6_mh_recv : udp4_mh_recv;
	int n, i;

	if ((n = udp_sock_recv(c, ref.fd, events, mmh_recv)) <= 0)
		return;

	/* We divide datagrams into batches based on how we need to send them,
	 * determined by udp_meta[i].tosidx.  To avoid either two passes through
	 * the array, or recalculating tosidx for a single entry, we have to
	 * populate it one entry *ahead* of the loop counter.
	 */
	udp_meta[0].tosidx = udp_flow_from_sock(c, ref, &udp_meta[0], now);
	for (i = 0; i < n; ) {
		flow_sidx_t batchsidx = udp_meta[i].tosidx;
		uint8_t batchpif = pif_at_sidx(batchsidx);
		int batchstart = i;

		do {
			if (pif_is_socket(batchpif)) {
				udp_splice_prepare(mmh_recv, i);
			} else if (batchpif == PIF_TAP) {
				udp_tap_prepare(mmh_recv, i,
						flowside_at_sidx(batchsidx));
			}

			if (++i >= n)
				break;

			udp_meta[i].tosidx = udp_flow_from_sock(c, ref,
								&udp_meta[i],
								now);
		} while (flow_sidx_eq(udp_meta[i].tosidx, batchsidx));

		if (pif_is_socket(batchpif)) {
			udp_splice_send(c, batchstart, i - batchstart,
					batchsidx);
		} else if (batchpif == PIF_TAP) {
			tap_send_frames(c, &udp_l2_iov[batchstart][0],
					UDP_NUM_IOVS, i - batchstart);
		} else if (flow_sidx_valid(batchsidx)) {
			flow_sidx_t fromsidx = flow_sidx_opposite(batchsidx);
			struct udp_flow *uflow = udp_at_sidx(batchsidx);

			flow_err(uflow,
				 "No support for forwarding UDP from %s to %s",
				 pif_name(pif_at_sidx(fromsidx)),
				 pif_name(batchpif));
		} else {
			debug("Discarding %d datagrams without flow",
			      i - batchstart);
		}
	}
}

/**
 * udp_reply_sock_handler() - Handle new data from flow specific socket
 * @c:		Execution context
 * @ref:	epoll reference
 * @events:	epoll events bitmap
 * @now:	Current timestamp
 *
 * #syscalls recvmmsg
 */
void udp_reply_sock_handler(const struct ctx *c, union epoll_ref ref,
			    uint32_t events, const struct timespec *now)
{
	const struct flowside *fromside = flowside_at_sidx(ref.flowside);
	flow_sidx_t tosidx = flow_sidx_opposite(ref.flowside);
	const struct flowside *toside = flowside_at_sidx(tosidx);
	struct udp_flow *uflow = udp_at_sidx(ref.flowside);
	int from_s = uflow->s[ref.flowside.sidei];
	bool v6 = !inany_v4(&fromside->eaddr);
	struct mmsghdr *mmh_recv = v6 ? udp6_mh_recv : udp4_mh_recv;
	uint8_t topif = pif_at_sidx(tosidx);
	int n, i;

	ASSERT(!c->no_udp && uflow);

	if ((n = udp_sock_recv(c, from_s, events, mmh_recv)) <= 0)
		return;

	flow_trace(uflow, "Received %d datagrams on reply socket", n);
	uflow->ts = now->tv_sec;

	for (i = 0; i < n; i++) {
		if (pif_is_socket(topif))
			udp_splice_prepare(mmh_recv, i);
		else if (topif == PIF_TAP)
			udp_tap_prepare(mmh_recv, i, toside);
	}

	if (pif_is_socket(topif)) {
		udp_splice_send(c, 0, n, tosidx);
	} else if (topif == PIF_TAP) {
		tap_send_frames(c, &udp_l2_iov[0][0], UDP_NUM_IOVS, n);
	} else {
		uint8_t frompif = pif_at_sidx(ref.flowside);

		flow_err(uflow, "No support for forwarding UDP from %s to %s",
			 pif_name(frompif), pif_name(topif));
	}
}

/**
 * udp_flow_from_tap() - Find or create UDP flow for tap packets
 * @c:		Execution context
 * @pif:	pif on which the packet is arriving
 * @af:		Address family, AF_INET or AF_INET6
 * @saddr:	Source address on guest side
 * @daddr:	Destination address guest side
 * @srcport:	Source port on guest side
 * @dstport:	Destination port on guest side
 *
 * Return: sidx for the destination side of the flow for this packet, or
 *         FLOW_SIDX_NONE if we couldn't find or create a flow.
 */
static flow_sidx_t udp_flow_from_tap(const struct ctx *c,
				     uint8_t pif, sa_family_t af,
				     const void *saddr, const void *daddr,
				     in_port_t srcport, in_port_t dstport,
				     const struct timespec *now)
{
	struct udp_flow *uflow;
	union flow *flow;
	flow_sidx_t sidx;

	ASSERT(pif == PIF_TAP);

	sidx = flow_lookup_af(c, IPPROTO_UDP, pif, af, saddr, daddr,
			      srcport, dstport);
	if ((uflow = udp_at_sidx(sidx))) {
		uflow->ts = now->tv_sec;
		return flow_sidx_opposite(sidx);
	}

	if (!(flow = flow_alloc())) {
		char sstr[INET6_ADDRSTRLEN], dstr[INET6_ADDRSTRLEN];

		debug("Couldn't allocate flow for UDP datagram from %s %s:%hu -> %s:%hu",
		      pif_name(pif),
		      inet_ntop(af, saddr, sstr, sizeof(sstr)), srcport,
		      inet_ntop(af, daddr, dstr, sizeof(dstr)), dstport);
		return FLOW_SIDX_NONE;
	}

	flow_initiate_af(flow, PIF_TAP, af, saddr, srcport, daddr, dstport);

	return udp_flow_new(c, flow, -1, now);
}

/**
 * udp_tap_handler() - Handle packets from tap
 * @c:		Execution context
 * @pif:	pif on which the packet is arriving
 * @af:		Address family, AF_INET or AF_INET6
 * @saddr:	Source address
 * @daddr:	Destination address
 * @p:		Pool of UDP packets, with UDP headers
 * @idx:	Index of first packet to process
 * @now:	Current timestamp
 *
 * Return: count of consumed packets
 *
 * #syscalls sendmmsg
 */
int udp_tap_handler(const struct ctx *c, uint8_t pif,
		    sa_family_t af, const void *saddr, const void *daddr,
		    const struct pool *p, int idx, const struct timespec *now)
{
	const struct flowside *toside;
	struct mmsghdr mm[UIO_MAXIOV];
	union sockaddr_inany to_sa;
	struct iovec m[UIO_MAXIOV];
	const struct udphdr *uh;
	struct udp_flow *uflow;
	int i, s, count = 0;
	flow_sidx_t tosidx;
	in_port_t src, dst;
	uint8_t topif;
	socklen_t sl;

	ASSERT(!c->no_udp);

	uh = packet_get(p, idx, 0, sizeof(*uh), NULL);
	if (!uh)
		return 1;

	/* The caller already checks that all the messages have the same source
	 * and destination, so we can just take those from the first message.
	 */
	src = ntohs(uh->source);
	dst = ntohs(uh->dest);

	tosidx = udp_flow_from_tap(c, pif, af, saddr, daddr, src, dst, now);
	if (!(uflow = udp_at_sidx(tosidx))) {
		char sstr[INET6_ADDRSTRLEN], dstr[INET6_ADDRSTRLEN];

		debug("Dropping datagram with no flow %s %s:%hu -> %s:%hu",
		      pif_name(pif),
		      inet_ntop(af, saddr, sstr, sizeof(sstr)), src,
		      inet_ntop(af, daddr, dstr, sizeof(dstr)), dst);
		return 1;
	}

	topif = pif_at_sidx(tosidx);
	if (topif != PIF_HOST) {
		flow_sidx_t fromsidx = flow_sidx_opposite(tosidx);
		uint8_t frompif = pif_at_sidx(fromsidx);

		flow_err(uflow, "No support for forwarding UDP from %s to %s",
			 pif_name(frompif), pif_name(topif));
		return 1;
	}
	toside = flowside_at_sidx(tosidx);

	s = udp_at_sidx(tosidx)->s[tosidx.sidei];
	ASSERT(s >= 0);

	pif_sockaddr(c, &to_sa, &sl, topif, &toside->eaddr, toside->eport);

	for (i = 0; i < (int)p->count - idx; i++) {
		struct udphdr *uh_send;
		size_t len;

		uh_send = packet_get(p, idx + i, 0, sizeof(*uh), &len);
		if (!uh_send)
			return p->count - idx;

		mm[i].msg_hdr.msg_name = &to_sa;
		mm[i].msg_hdr.msg_namelen = sl;

		if (len) {
			m[i].iov_base = (char *)(uh_send + 1);
			m[i].iov_len = len;

			mm[i].msg_hdr.msg_iov = m + i;
			mm[i].msg_hdr.msg_iovlen = 1;
		} else {
			mm[i].msg_hdr.msg_iov = NULL;
			mm[i].msg_hdr.msg_iovlen = 0;
		}

		mm[i].msg_hdr.msg_control = NULL;
		mm[i].msg_hdr.msg_controllen = 0;
		mm[i].msg_hdr.msg_flags = 0;

		count++;
	}

	count = sendmmsg(s, mm, count, MSG_NOSIGNAL);
	if (count < 0)
		return 1;

	return count;
}

/**
 * udp_sock_init() - Initialise listening sockets for a given port
 * @c:		Execution context
 * @ns:		In pasta mode, if set, bind with loopback address in namespace
 * @af:		Address family to select a specific IP version, or AF_UNSPEC
 * @addr:	Pointer to address for binding, NULL if not configured
 * @ifname:	Name of interface to bind to, NULL if not configured
 * @port:	Port, host order
 *
 * Return: 0 on (partial) success, negative error code on (complete) failure
 */
int udp_sock_init(const struct ctx *c, int ns, sa_family_t af,
		  const void *addr, const char *ifname, in_port_t port)
{
	union udp_epoll_ref uref = { .orig = true, .port = port };
	int s, r4 = FD_REF_MAX + 1, r6 = FD_REF_MAX + 1;

	ASSERT(!c->no_udp);

	if (ns)
		uref.pif = PIF_SPLICE;
	else
		uref.pif = PIF_HOST;

	if ((af == AF_INET || af == AF_UNSPEC) && c->ifi4) {
		uref.v6 = 0;

		if (!ns) {
			r4 = s = sock_l4(c, AF_INET, EPOLL_TYPE_UDP, addr,
					 ifname, port, uref.u32);

			udp_splice_init[V4][port] = s < 0 ? -1 : s;
		} else {
			r4 = s = sock_l4(c, AF_INET, EPOLL_TYPE_UDP,
					 &in4addr_loopback,
					 ifname, port, uref.u32);
			udp_splice_ns[V4][port] = s < 0 ? -1 : s;
		}
	}

	if ((af == AF_INET6 || af == AF_UNSPEC) && c->ifi6) {
		uref.v6 = 1;

		if (!ns) {
			r6 = s = sock_l4(c, AF_INET6, EPOLL_TYPE_UDP, addr,
					 ifname, port, uref.u32);

			udp_splice_init[V6][port] = s < 0 ? -1 : s;
		} else {
			r6 = s = sock_l4(c, AF_INET6, EPOLL_TYPE_UDP,
					 &in6addr_loopback,
					 ifname, port, uref.u32);
			udp_splice_ns[V6][port] = s < 0 ? -1 : s;
		}
	}

	if (IN_INTERVAL(0, FD_REF_MAX, r4) || IN_INTERVAL(0, FD_REF_MAX, r6))
		return 0;

	return r4 < 0 ? r4 : r6;
}

/**
 * udp_splice_iov_init() - Set up buffers and descriptors for recvmmsg/sendmmsg
 */
static void udp_splice_iov_init(void)
{
	int i;

	for (i = 0; i < UDP_MAX_FRAMES; i++) {
		struct msghdr *mh = &udp_mh_splice[i].msg_hdr;

		mh->msg_name = &udp_splice_to;
		mh->msg_namelen = sizeof(udp_splice_to);

		udp_iov_splice[i].iov_base = udp_payload[i].data;

		mh->msg_iov = &udp_iov_splice[i];
		mh->msg_iovlen = 1;
	}
}

/**
 * udp_port_rebind() - Rebind ports to match forward maps
 * @c:		Execution context
 * @outbound:	True to remap outbound forwards, otherwise inbound
 *
 * Must be called in namespace context if @outbound is true.
 */
static void udp_port_rebind(struct ctx *c, bool outbound)
{
	int (*socks)[NUM_PORTS] = outbound ? udp_splice_ns : udp_splice_init;
	const uint8_t *fmap
		= outbound ? c->udp.fwd_out.map : c->udp.fwd_in.map;
	const uint8_t *rmap
		= outbound ? c->udp.fwd_in.map : c->udp.fwd_out.map;
	unsigned port;

	for (port = 0; port < NUM_PORTS; port++) {
		if (!bitmap_isset(fmap, port)) {
			if (socks[V4][port] >= 0) {
				close(socks[V4][port]);
				socks[V4][port] = -1;
			}

			if (socks[V6][port] >= 0) {
				close(socks[V6][port]);
				socks[V6][port] = -1;
			}

			continue;
		}

		/* Don't loop back our own ports */
		if (bitmap_isset(rmap, port))
			continue;

		if ((c->ifi4 && socks[V4][port] == -1) ||
		    (c->ifi6 && socks[V6][port] == -1))
			udp_sock_init(c, outbound, AF_UNSPEC, NULL, NULL, port);
	}
}

/**
 * udp_port_rebind_outbound() - Rebind ports in namespace
 * @arg:	Execution context
 *
 * Called with NS_CALL()
 *
 * Return: 0
 */
static int udp_port_rebind_outbound(void *arg)
{
	struct ctx *c = (struct ctx *)arg;

	ns_enter(c);
	udp_port_rebind(c, true);

	return 0;
}

/**
 * udp_flow_timer() - Handler for timed events related to a given flow
 * @c:		Execution context
 * @uflow:	UDP flow
 * @now:	Current timestamp
 *
 * Return: true if the flow is ready to free, false otherwise
 */
bool udp_flow_timer(const struct ctx *c, struct udp_flow *uflow,
		    const struct timespec *now)
{
	if (now->tv_sec - uflow->ts <= UDP_CONN_TIMEOUT)
		return false;

	udp_flow_close(c, uflow);
	return true;
}

/**
 * udp_timer() - Scan activity bitmaps for ports with associated timed events
 * @c:		Execution context
 * @now:	Current timestamp
 */
void udp_timer(struct ctx *c, const struct timespec *now)
{
	(void)now;

	ASSERT(!c->no_udp);

	if (c->mode == MODE_PASTA) {
		if (c->udp.fwd_out.mode == FWD_AUTO) {
			fwd_scan_ports_udp(&c->udp.fwd_out, &c->udp.fwd_in,
					   &c->tcp.fwd_out, &c->tcp.fwd_in);
			NS_CALL(udp_port_rebind_outbound, c);
		}

		if (c->udp.fwd_in.mode == FWD_AUTO) {
			fwd_scan_ports_udp(&c->udp.fwd_in, &c->udp.fwd_out,
					   &c->tcp.fwd_in, &c->tcp.fwd_out);
			udp_port_rebind(c, false);
		}
	}
}

/**
 * udp_init() - Initialise per-socket data, and sockets in namespace
 * @c:		Execution context
 *
 * Return: 0
 */
int udp_init(struct ctx *c)
{
	ASSERT(!c->no_udp);

	udp_iov_init(c);

	if (c->mode == MODE_PASTA) {
		udp_splice_iov_init();
		NS_CALL(udp_port_rebind_outbound, c);
	}

	return 0;
}