/
netstack.go
3438 lines (2955 loc) · 113 KB
/
netstack.go
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
// Copyright 2018 The gVisor Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
// Package netstack provides an implementation of the socket.Socket interface
// that is backed by a tcpip.Endpoint.
//
// It does not depend on any particular endpoint implementation, and thus can
// be used to expose certain endpoints to the sentry while leaving others out,
// for example, TCP endpoints and Unix-domain endpoints.
//
// Lock ordering: netstack => mm: ioSequenceReadWriter copies user memory inside
// tcpip.Endpoint.Write(). Netstack is allowed to (and does) hold locks during
// this operation.
package netstack
import (
"bytes"
"encoding/binary"
"fmt"
"io"
"io/ioutil"
"math"
"reflect"
"time"
"golang.org/x/sys/unix"
"google.golang.org/protobuf/proto"
"gvisor.dev/gvisor/pkg/abi/linux"
"gvisor.dev/gvisor/pkg/abi/linux/errno"
"gvisor.dev/gvisor/pkg/context"
"gvisor.dev/gvisor/pkg/errors/linuxerr"
"gvisor.dev/gvisor/pkg/eventchannel"
"gvisor.dev/gvisor/pkg/hostarch"
"gvisor.dev/gvisor/pkg/log"
"gvisor.dev/gvisor/pkg/marshal"
"gvisor.dev/gvisor/pkg/marshal/primitive"
"gvisor.dev/gvisor/pkg/metric"
"gvisor.dev/gvisor/pkg/sentry/arch"
"gvisor.dev/gvisor/pkg/sentry/fsimpl/sockfs"
"gvisor.dev/gvisor/pkg/sentry/inet"
"gvisor.dev/gvisor/pkg/sentry/kernel"
"gvisor.dev/gvisor/pkg/sentry/kernel/auth"
ktime "gvisor.dev/gvisor/pkg/sentry/kernel/time"
"gvisor.dev/gvisor/pkg/sentry/socket"
"gvisor.dev/gvisor/pkg/sentry/socket/netfilter"
epb "gvisor.dev/gvisor/pkg/sentry/socket/netstack/events_go_proto"
"gvisor.dev/gvisor/pkg/sentry/vfs"
"gvisor.dev/gvisor/pkg/sync"
"gvisor.dev/gvisor/pkg/syserr"
"gvisor.dev/gvisor/pkg/tcpip"
"gvisor.dev/gvisor/pkg/tcpip/header"
"gvisor.dev/gvisor/pkg/tcpip/stack"
"gvisor.dev/gvisor/pkg/tcpip/transport"
"gvisor.dev/gvisor/pkg/tcpip/transport/tcp"
"gvisor.dev/gvisor/pkg/usermem"
"gvisor.dev/gvisor/pkg/waiter"
)
const bitsPerUint32 = 32
// statCounterValue returns a function usable as callback function when defining a gVisor Sentry
// metric that contains the value counted by the StatCounter.
// This avoids a dependency loop in the tcpip package.
func statCounterValue(cm *tcpip.StatCounter) func(...*metric.FieldValue) uint64 {
return func(...*metric.FieldValue) uint64 {
return cm.Value()
}
}
func mustCreateMetric(name, description string) *tcpip.StatCounter {
var cm tcpip.StatCounter
metric.MustRegisterCustomUint64Metric(name, true /* cumulative */, false /* sync */, description, statCounterValue(&cm))
return &cm
}
func mustCreateGauge(name, description string) *tcpip.StatCounter {
var cm tcpip.StatCounter
metric.MustRegisterCustomUint64Metric(name, false /* cumulative */, false /* sync */, description, statCounterValue(&cm))
return &cm
}
// Metrics contains metrics exported by netstack.
var Metrics = tcpip.Stats{
DroppedPackets: mustCreateMetric("/netstack/dropped_packets", "Number of packets dropped at the transport layer."),
NICs: tcpip.NICStats{
MalformedL4RcvdPackets: mustCreateMetric("/netstack/nic/malformed_l4_received_packets", "Number of packets received that failed L4 header parsing."),
Tx: tcpip.NICPacketStats{
Packets: mustCreateMetric("/netstack/nic/tx/packets", "Number of packets transmitted."),
Bytes: mustCreateMetric("/netstack/nic/tx/bytes", "Number of bytes transmitted."),
},
TxPacketsDroppedNoBufferSpace: mustCreateMetric("/netstack/nic/tx_packets_dropped_no_buffer_space", "Number of TX packets dropped as a result of no buffer space errors."),
Rx: tcpip.NICPacketStats{
Packets: mustCreateMetric("/netstack/nic/rx/packets", "Number of packets received."),
Bytes: mustCreateMetric("/netstack/nic/rx/bytes", "Number of bytes received."),
},
DisabledRx: tcpip.NICPacketStats{
Packets: mustCreateMetric("/netstack/nic/disabled_rx/packets", "Number of packets received on disabled NICs."),
Bytes: mustCreateMetric("/netstack/nic/disabled_rx/bytes", "Number of bytes received on disabled NICs."),
},
Neighbor: tcpip.NICNeighborStats{
UnreachableEntryLookups: mustCreateMetric("/netstack/nic/neighbor/unreachable_entry_loopups", "Number of lookups performed on a neighbor entry in Unreachable state."),
DroppedConfirmationForNoninitiatedNeighbor: mustCreateMetric("/netstack/nic/neighbor/dropped_confirmation_for_noninitiated_neighbor", "Number of advertisements received that don't match an entry in the neighbor cache."),
DroppedInvalidLinkAddressConfirmations: mustCreateMetric("/netstack/nic/neighbor/dropped_invalid_link_address_confirmations", "Number of advertisements dropped because they have empty source link-layer addresses"),
},
},
ICMP: tcpip.ICMPStats{
V4: tcpip.ICMPv4Stats{
PacketsSent: tcpip.ICMPv4SentPacketStats{
ICMPv4PacketStats: tcpip.ICMPv4PacketStats{
EchoRequest: mustCreateMetric("/netstack/icmp/v4/packets_sent/echo_request", "Number of ICMPv4 echo request packets sent."),
EchoReply: mustCreateMetric("/netstack/icmp/v4/packets_sent/echo_reply", "Number of ICMPv4 echo reply packets sent."),
DstUnreachable: mustCreateMetric("/netstack/icmp/v4/packets_sent/dst_unreachable", "Number of ICMPv4 destination unreachable packets sent."),
SrcQuench: mustCreateMetric("/netstack/icmp/v4/packets_sent/src_quench", "Number of ICMPv4 source quench packets sent."),
Redirect: mustCreateMetric("/netstack/icmp/v4/packets_sent/redirect", "Number of ICMPv4 redirect packets sent."),
TimeExceeded: mustCreateMetric("/netstack/icmp/v4/packets_sent/time_exceeded", "Number of ICMPv4 time exceeded packets sent."),
ParamProblem: mustCreateMetric("/netstack/icmp/v4/packets_sent/param_problem", "Number of ICMPv4 parameter problem packets sent."),
Timestamp: mustCreateMetric("/netstack/icmp/v4/packets_sent/timestamp", "Number of ICMPv4 timestamp packets sent."),
TimestampReply: mustCreateMetric("/netstack/icmp/v4/packets_sent/timestamp_reply", "Number of ICMPv4 timestamp reply packets sent."),
InfoRequest: mustCreateMetric("/netstack/icmp/v4/packets_sent/info_request", "Number of ICMPv4 information request packets sent."),
InfoReply: mustCreateMetric("/netstack/icmp/v4/packets_sent/info_reply", "Number of ICMPv4 information reply packets sent."),
},
Dropped: mustCreateMetric("/netstack/icmp/v4/packets_sent/dropped", "Number of ICMPv4 packets dropped due to link layer errors."),
RateLimited: mustCreateMetric("/netstack/icmp/v4/packets_sent/rate_limited", "Number of ICMPv4 packets dropped due to rate limit being exceeded."),
},
PacketsReceived: tcpip.ICMPv4ReceivedPacketStats{
ICMPv4PacketStats: tcpip.ICMPv4PacketStats{
EchoRequest: mustCreateMetric("/netstack/icmp/v4/packets_received/echo_request", "Number of ICMPv4 echo request packets received."),
EchoReply: mustCreateMetric("/netstack/icmp/v4/packets_received/echo_reply", "Number of ICMPv4 echo reply packets received."),
DstUnreachable: mustCreateMetric("/netstack/icmp/v4/packets_received/dst_unreachable", "Number of ICMPv4 destination unreachable packets received."),
SrcQuench: mustCreateMetric("/netstack/icmp/v4/packets_received/src_quench", "Number of ICMPv4 source quench packets received."),
Redirect: mustCreateMetric("/netstack/icmp/v4/packets_received/redirect", "Number of ICMPv4 redirect packets received."),
TimeExceeded: mustCreateMetric("/netstack/icmp/v4/packets_received/time_exceeded", "Number of ICMPv4 time exceeded packets received."),
ParamProblem: mustCreateMetric("/netstack/icmp/v4/packets_received/param_problem", "Number of ICMPv4 parameter problem packets received."),
Timestamp: mustCreateMetric("/netstack/icmp/v4/packets_received/timestamp", "Number of ICMPv4 timestamp packets received."),
TimestampReply: mustCreateMetric("/netstack/icmp/v4/packets_received/timestamp_reply", "Number of ICMPv4 timestamp reply packets received."),
InfoRequest: mustCreateMetric("/netstack/icmp/v4/packets_received/info_request", "Number of ICMPv4 information request packets received."),
InfoReply: mustCreateMetric("/netstack/icmp/v4/packets_received/info_reply", "Number of ICMPv4 information reply packets received."),
},
Invalid: mustCreateMetric("/netstack/icmp/v4/packets_received/invalid", "Number of ICMPv4 packets received that the transport layer could not parse."),
},
},
V6: tcpip.ICMPv6Stats{
PacketsSent: tcpip.ICMPv6SentPacketStats{
ICMPv6PacketStats: tcpip.ICMPv6PacketStats{
EchoRequest: mustCreateMetric("/netstack/icmp/v6/packets_sent/echo_request", "Number of ICMPv6 echo request packets sent."),
EchoReply: mustCreateMetric("/netstack/icmp/v6/packets_sent/echo_reply", "Number of ICMPv6 echo reply packets sent."),
DstUnreachable: mustCreateMetric("/netstack/icmp/v6/packets_sent/dst_unreachable", "Number of ICMPv6 destination unreachable packets sent."),
PacketTooBig: mustCreateMetric("/netstack/icmp/v6/packets_sent/packet_too_big", "Number of ICMPv6 packet too big packets sent."),
TimeExceeded: mustCreateMetric("/netstack/icmp/v6/packets_sent/time_exceeded", "Number of ICMPv6 time exceeded packets sent."),
ParamProblem: mustCreateMetric("/netstack/icmp/v6/packets_sent/param_problem", "Number of ICMPv6 parameter problem packets sent."),
RouterSolicit: mustCreateMetric("/netstack/icmp/v6/packets_sent/router_solicit", "Number of ICMPv6 router solicit packets sent."),
RouterAdvert: mustCreateMetric("/netstack/icmp/v6/packets_sent/router_advert", "Number of ICMPv6 router advert packets sent."),
NeighborSolicit: mustCreateMetric("/netstack/icmp/v6/packets_sent/neighbor_solicit", "Number of ICMPv6 neighbor solicit packets sent."),
NeighborAdvert: mustCreateMetric("/netstack/icmp/v6/packets_sent/neighbor_advert", "Number of ICMPv6 neighbor advert packets sent."),
RedirectMsg: mustCreateMetric("/netstack/icmp/v6/packets_sent/redirect_msg", "Number of ICMPv6 redirect message packets sent."),
MulticastListenerQuery: mustCreateMetric("/netstack/icmp/v6/packets_sent/multicast_listener_query", "Number of ICMPv6 multicast listener query packets sent."),
MulticastListenerReport: mustCreateMetric("/netstack/icmp/v6/packets_sent/multicast_listener_report", "Number of ICMPv6 multicast listener report packets sent."),
MulticastListenerDone: mustCreateMetric("/netstack/icmp/v6/packets_sent/multicast_listener_done", "Number of ICMPv6 multicast listener done packets sent."),
},
Dropped: mustCreateMetric("/netstack/icmp/v6/packets_sent/dropped", "Number of ICMPv6 packets dropped due to link layer errors."),
RateLimited: mustCreateMetric("/netstack/icmp/v6/packets_sent/rate_limited", "Number of ICMPv6 packets dropped due to rate limit being exceeded."),
},
PacketsReceived: tcpip.ICMPv6ReceivedPacketStats{
ICMPv6PacketStats: tcpip.ICMPv6PacketStats{
EchoRequest: mustCreateMetric("/netstack/icmp/v6/packets_received/echo_request", "Number of ICMPv6 echo request packets received."),
EchoReply: mustCreateMetric("/netstack/icmp/v6/packets_received/echo_reply", "Number of ICMPv6 echo reply packets received."),
DstUnreachable: mustCreateMetric("/netstack/icmp/v6/packets_received/dst_unreachable", "Number of ICMPv6 destination unreachable packets received."),
PacketTooBig: mustCreateMetric("/netstack/icmp/v6/packets_received/packet_too_big", "Number of ICMPv6 packet too big packets received."),
TimeExceeded: mustCreateMetric("/netstack/icmp/v6/packets_received/time_exceeded", "Number of ICMPv6 time exceeded packets received."),
ParamProblem: mustCreateMetric("/netstack/icmp/v6/packets_received/param_problem", "Number of ICMPv6 parameter problem packets received."),
RouterSolicit: mustCreateMetric("/netstack/icmp/v6/packets_received/router_solicit", "Number of ICMPv6 router solicit packets received."),
RouterAdvert: mustCreateMetric("/netstack/icmp/v6/packets_received/router_advert", "Number of ICMPv6 router advert packets received."),
NeighborSolicit: mustCreateMetric("/netstack/icmp/v6/packets_received/neighbor_solicit", "Number of ICMPv6 neighbor solicit packets received."),
NeighborAdvert: mustCreateMetric("/netstack/icmp/v6/packets_received/neighbor_advert", "Number of ICMPv6 neighbor advert packets received."),
RedirectMsg: mustCreateMetric("/netstack/icmp/v6/packets_received/redirect_msg", "Number of ICMPv6 redirect message packets received."),
MulticastListenerQuery: mustCreateMetric("/netstack/icmp/v6/packets_received/multicast_listener_query", "Number of ICMPv6 multicast listener query packets received."),
MulticastListenerReport: mustCreateMetric("/netstack/icmp/v6/packets_received/multicast_listener_report", "Number of ICMPv6 multicast listener report packets sent."),
MulticastListenerDone: mustCreateMetric("/netstack/icmp/v6/packets_received/multicast_listener_done", "Number of ICMPv6 multicast listener done packets sent."),
},
Unrecognized: mustCreateMetric("/netstack/icmp/v6/packets_received/unrecognized", "Number of ICMPv6 packets received that the transport layer does not know how to parse."),
Invalid: mustCreateMetric("/netstack/icmp/v6/packets_received/invalid", "Number of ICMPv6 packets received that the transport layer could not parse."),
RouterOnlyPacketsDroppedByHost: mustCreateMetric("/netstack/icmp/v6/packets_received/router_only_packets_dropped_by_host", "Number of ICMPv6 packets dropped due to being router-specific packets."),
},
},
},
IGMP: tcpip.IGMPStats{
PacketsSent: tcpip.IGMPSentPacketStats{
IGMPPacketStats: tcpip.IGMPPacketStats{
MembershipQuery: mustCreateMetric("/netstack/igmp/packets_sent/membership_query", "Number of IGMP Membership Query messages sent."),
V1MembershipReport: mustCreateMetric("/netstack/igmp/packets_sent/v1_membership_report", "Number of IGMPv1 Membership Report messages sent."),
V2MembershipReport: mustCreateMetric("/netstack/igmp/packets_sent/v2_membership_report", "Number of IGMPv2 Membership Report messages sent."),
LeaveGroup: mustCreateMetric("/netstack/igmp/packets_sent/leave_group", "Number of IGMP Leave Group messages sent."),
},
Dropped: mustCreateMetric("/netstack/igmp/packets_sent/dropped", "Number of IGMP packets dropped due to link layer errors."),
},
PacketsReceived: tcpip.IGMPReceivedPacketStats{
IGMPPacketStats: tcpip.IGMPPacketStats{
MembershipQuery: mustCreateMetric("/netstack/igmp/packets_received/membership_query", "Number of IGMP Membership Query messages received."),
V1MembershipReport: mustCreateMetric("/netstack/igmp/packets_received/v1_membership_report", "Number of IGMPv1 Membership Report messages received."),
V2MembershipReport: mustCreateMetric("/netstack/igmp/packets_received/v2_membership_report", "Number of IGMPv2 Membership Report messages received."),
LeaveGroup: mustCreateMetric("/netstack/igmp/packets_received/leave_group", "Number of IGMP Leave Group messages received."),
},
Invalid: mustCreateMetric("/netstack/igmp/packets_received/invalid", "Number of IGMP packets received that could not be parsed."),
ChecksumErrors: mustCreateMetric("/netstack/igmp/packets_received/checksum_errors", "Number of received IGMP packets with bad checksums."),
Unrecognized: mustCreateMetric("/netstack/igmp/packets_received/unrecognized", "Number of unrecognized IGMP packets received."),
},
},
IP: tcpip.IPStats{
PacketsReceived: mustCreateMetric("/netstack/ip/packets_received", "Number of IP packets received from the link layer in nic.DeliverNetworkPacket."),
DisabledPacketsReceived: mustCreateMetric("/netstack/ip/disabled_packets_received", "Number of IP packets received from the link layer when the IP layer is disabled."),
InvalidDestinationAddressesReceived: mustCreateMetric("/netstack/ip/invalid_addresses_received", "Number of IP packets received with an unknown or invalid destination address."),
InvalidSourceAddressesReceived: mustCreateMetric("/netstack/ip/invalid_source_addresses_received", "Number of IP packets received with an unknown or invalid source address."),
PacketsDelivered: mustCreateMetric("/netstack/ip/packets_delivered", "Number of incoming IP packets that are successfully delivered to the transport layer via HandlePacket."),
PacketsSent: mustCreateMetric("/netstack/ip/packets_sent", "Number of IP packets sent via WritePacket."),
OutgoingPacketErrors: mustCreateMetric("/netstack/ip/outgoing_packet_errors", "Number of IP packets which failed to write to a link-layer endpoint."),
MalformedPacketsReceived: mustCreateMetric("/netstack/ip/malformed_packets_received", "Number of IP packets which failed IP header validation checks."),
MalformedFragmentsReceived: mustCreateMetric("/netstack/ip/malformed_fragments_received", "Number of IP fragments which failed IP fragment validation checks."),
IPTablesPreroutingDropped: mustCreateMetric("/netstack/ip/iptables/prerouting_dropped", "Number of IP packets dropped in the Prerouting chain."),
IPTablesInputDropped: mustCreateMetric("/netstack/ip/iptables/input_dropped", "Number of IP packets dropped in the Input chain."),
IPTablesOutputDropped: mustCreateMetric("/netstack/ip/iptables/output_dropped", "Number of IP packets dropped in the Output chain."),
OptionTimestampReceived: mustCreateMetric("/netstack/ip/options/timestamp_received", "Number of timestamp options found in received IP packets."),
OptionRecordRouteReceived: mustCreateMetric("/netstack/ip/options/record_route_received", "Number of record route options found in received IP packets."),
OptionRouterAlertReceived: mustCreateMetric("/netstack/ip/options/router_alert_received", "Number of router alert options found in received IP packets."),
OptionUnknownReceived: mustCreateMetric("/netstack/ip/options/unknown_received", "Number of unknown options found in received IP packets."),
Forwarding: tcpip.IPForwardingStats{
Unrouteable: mustCreateMetric("/netstack/ip/forwarding/unrouteable", "Number of IP packets received which couldn't be routed and thus were not forwarded."),
ExhaustedTTL: mustCreateMetric("/netstack/ip/forwarding/exhausted_ttl", "Number of IP packets received which could not be forwarded due to an exhausted TTL."),
LinkLocalSource: mustCreateMetric("/netstack/ip/forwarding/link_local_source_address", "Number of IP packets received which could not be forwarded due to a link-local source address."),
LinkLocalDestination: mustCreateMetric("/netstack/ip/forwarding/link_local_destination_address", "Number of IP packets received which could not be forwarded due to a link-local destination address."),
ExtensionHeaderProblem: mustCreateMetric("/netstack/ip/forwarding/extension_header_problem", "Number of IP packets received which could not be forwarded due to a problem processing their IPv6 extension headers."),
PacketTooBig: mustCreateMetric("/netstack/ip/forwarding/packet_too_big", "Number of IP packets received which could not be forwarded because they could not fit within the outgoing MTU."),
HostUnreachable: mustCreateMetric("/netstack/ip/forwarding/host_unreachable", "Number of IP packets received which could not be forwarded due to unresolvable next hop."),
Errors: mustCreateMetric("/netstack/ip/forwarding/errors", "Number of IP packets which couldn't be forwarded."),
},
},
ARP: tcpip.ARPStats{
PacketsReceived: mustCreateMetric("/netstack/arp/packets_received", "Number of ARP packets received from the link layer."),
DisabledPacketsReceived: mustCreateMetric("/netstack/arp/disabled_packets_received", "Number of ARP packets received from the link layer when the ARP layer is disabled."),
MalformedPacketsReceived: mustCreateMetric("/netstack/arp/malformed_packets_received", "Number of ARP packets which failed ARP header validation checks."),
RequestsReceived: mustCreateMetric("/netstack/arp/requests_received", "Number of ARP requests received."),
RequestsReceivedUnknownTargetAddress: mustCreateMetric("/netstack/arp/requests_received_unknown_addr", "Number of ARP requests received with an unknown target address."),
OutgoingRequestInterfaceHasNoLocalAddressErrors: mustCreateMetric("/netstack/arp/outgoing_requests_iface_has_no_addr", "Number of failed attempts to send an ARP request with an interface that has no network address."),
OutgoingRequestBadLocalAddressErrors: mustCreateMetric("/netstack/arp/outgoing_requests_invalid_local_addr", "Number of failed attempts to send an ARP request with a provided local address that is invalid."),
OutgoingRequestsDropped: mustCreateMetric("/netstack/arp/outgoing_requests_dropped", "Number of ARP requests which failed to write to a link-layer endpoint."),
OutgoingRequestsSent: mustCreateMetric("/netstack/arp/outgoing_requests_sent", "Number of ARP requests sent."),
RepliesReceived: mustCreateMetric("/netstack/arp/replies_received", "Number of ARP replies received."),
OutgoingRepliesDropped: mustCreateMetric("/netstack/arp/outgoing_replies_dropped", "Number of ARP replies which failed to write to a link-layer endpoint."),
OutgoingRepliesSent: mustCreateMetric("/netstack/arp/outgoing_replies_sent", "Number of ARP replies sent."),
},
TCP: tcpip.TCPStats{
ActiveConnectionOpenings: mustCreateMetric("/netstack/tcp/active_connection_openings", "Number of connections opened successfully via Connect."),
PassiveConnectionOpenings: mustCreateMetric("/netstack/tcp/passive_connection_openings", "Number of connections opened successfully via Listen."),
CurrentEstablished: mustCreateGauge("/netstack/tcp/current_established", "Number of connections in ESTABLISHED state now."),
CurrentConnected: mustCreateGauge("/netstack/tcp/current_open", "Number of connections that are in connected state."),
EstablishedResets: mustCreateMetric("/netstack/tcp/established_resets", "Number of times TCP connections have made a direct transition to the CLOSED state from either the ESTABLISHED state or the CLOSE-WAIT state"),
EstablishedClosed: mustCreateMetric("/netstack/tcp/established_closed", "Number of times established TCP connections made a transition to CLOSED state."),
EstablishedTimedout: mustCreateMetric("/netstack/tcp/established_timedout", "Number of times an established connection was reset because of keep-alive time out."),
ListenOverflowSynDrop: mustCreateMetric("/netstack/tcp/listen_overflow_syn_drop", "Number of times the listen queue overflowed and a SYN was dropped."),
ListenOverflowAckDrop: mustCreateMetric("/netstack/tcp/listen_overflow_ack_drop", "Number of times the listen queue overflowed and the final ACK in the handshake was dropped."),
ListenOverflowSynCookieSent: mustCreateMetric("/netstack/tcp/listen_overflow_syn_cookie_sent", "Number of times a SYN cookie was sent."),
ListenOverflowSynCookieRcvd: mustCreateMetric("/netstack/tcp/listen_overflow_syn_cookie_rcvd", "Number of times a SYN cookie was received."),
ListenOverflowInvalidSynCookieRcvd: mustCreateMetric("/netstack/tcp/listen_overflow_invalid_syn_cookie_rcvd", "Number of times an invalid SYN cookie was received."),
FailedConnectionAttempts: mustCreateMetric("/netstack/tcp/failed_connection_attempts", "Number of calls to Connect or Listen (active and passive openings, respectively) that end in an error."),
ValidSegmentsReceived: mustCreateMetric("/netstack/tcp/valid_segments_received", "Number of TCP segments received that the transport layer successfully parsed."),
InvalidSegmentsReceived: mustCreateMetric("/netstack/tcp/invalid_segments_received", "Number of TCP segments received that the transport layer could not parse."),
SegmentsSent: mustCreateMetric("/netstack/tcp/segments_sent", "Number of TCP segments sent."),
SegmentSendErrors: mustCreateMetric("/netstack/tcp/segment_send_errors", "Number of TCP segments failed to be sent."),
ResetsSent: mustCreateMetric("/netstack/tcp/resets_sent", "Number of TCP resets sent."),
ResetsReceived: mustCreateMetric("/netstack/tcp/resets_received", "Number of TCP resets received."),
Retransmits: mustCreateMetric("/netstack/tcp/retransmits", "Number of TCP segments retransmitted."),
FastRecovery: mustCreateMetric("/netstack/tcp/fast_recovery", "Number of times fast recovery was used to recover from packet loss."),
SACKRecovery: mustCreateMetric("/netstack/tcp/sack_recovery", "Number of times SACK recovery was used to recover from packet loss."),
TLPRecovery: mustCreateMetric("/netstack/tcp/tlp_recovery", "Number of times tail loss probe triggers recovery from tail loss."),
SlowStartRetransmits: mustCreateMetric("/netstack/tcp/slow_start_retransmits", "Number of segments retransmitted in slow start mode."),
FastRetransmit: mustCreateMetric("/netstack/tcp/fast_retransmit", "Number of TCP segments which were fast retransmitted."),
Timeouts: mustCreateMetric("/netstack/tcp/timeouts", "Number of times RTO expired."),
ChecksumErrors: mustCreateMetric("/netstack/tcp/checksum_errors", "Number of segments dropped due to bad checksums."),
FailedPortReservations: mustCreateMetric("/netstack/tcp/failed_port_reservations", "Number of time TCP failed to reserve a port."),
SegmentsAckedWithDSACK: mustCreateMetric("/netstack/tcp/segments_acked_with_dsack", "Number of segments for which DSACK was received."),
SpuriousRecovery: mustCreateMetric("/netstack/tcp/spurious_recovery", "Number of times the connection entered loss recovery spuriously."),
SpuriousRTORecovery: mustCreateMetric("/netstack/tcp/spurious_rto_recovery", "Number of times the connection entered RTO spuriously."),
ForwardMaxInFlightDrop: mustCreateMetric("/netstack/tcp/forward_max_in_flight_drop", "Number of connection requests dropped due to exceeding in-flight limit."),
},
UDP: tcpip.UDPStats{
PacketsReceived: mustCreateMetric("/netstack/udp/packets_received", "Number of UDP datagrams received via HandlePacket."),
UnknownPortErrors: mustCreateMetric("/netstack/udp/unknown_port_errors", "Number of incoming UDP datagrams dropped because they did not have a known destination port."),
ReceiveBufferErrors: mustCreateMetric("/netstack/udp/receive_buffer_errors", "Number of incoming UDP datagrams dropped due to the receiving buffer being in an invalid state."),
MalformedPacketsReceived: mustCreateMetric("/netstack/udp/malformed_packets_received", "Number of incoming UDP datagrams dropped due to the UDP header being in a malformed state."),
PacketsSent: mustCreateMetric("/netstack/udp/packets_sent", "Number of UDP datagrams sent."),
PacketSendErrors: mustCreateMetric("/netstack/udp/packet_send_errors", "Number of UDP datagrams failed to be sent."),
ChecksumErrors: mustCreateMetric("/netstack/udp/checksum_errors", "Number of UDP datagrams dropped due to bad checksums."),
},
}
// DefaultTTL is linux's default TTL. All network protocols in all stacks used
// with this package must have this value set as their default TTL.
const DefaultTTL = 64
const sizeOfInt32 int = 4
var errStackType = syserr.New("expected but did not receive a netstack.Stack", errno.EINVAL)
// commonEndpoint represents the intersection of a tcpip.Endpoint and a
// transport.Endpoint.
type commonEndpoint interface {
// Readiness implements tcpip.Endpoint.Readiness and
// transport.Endpoint.Readiness.
Readiness(mask waiter.EventMask) waiter.EventMask
// SetSockOpt implements tcpip.Endpoint.SetSockOpt and
// transport.Endpoint.SetSockOpt.
SetSockOpt(tcpip.SettableSocketOption) tcpip.Error
// SetSockOptInt implements tcpip.Endpoint.SetSockOptInt and
// transport.Endpoint.SetSockOptInt.
SetSockOptInt(opt tcpip.SockOptInt, v int) tcpip.Error
// GetSockOpt implements tcpip.Endpoint.GetSockOpt and
// transport.Endpoint.GetSockOpt.
GetSockOpt(tcpip.GettableSocketOption) tcpip.Error
// GetSockOptInt implements tcpip.Endpoint.GetSockOptInt and
// transport.Endpoint.GetSockOpt.
GetSockOptInt(opt tcpip.SockOptInt) (int, tcpip.Error)
// State returns a socket's lifecycle state. The returned value is
// protocol-specific and is primarily used for diagnostics.
State() uint32
// LastError implements tcpip.Endpoint.LastError and
// transport.Endpoint.LastError.
LastError() tcpip.Error
// SocketOptions implements tcpip.Endpoint.SocketOptions and
// transport.Endpoint.SocketOptions.
SocketOptions() *tcpip.SocketOptions
}
// sock encapsulates all the state needed to represent a network stack
// endpoint in the kernel context.
//
// +stateify savable
type sock struct {
vfsfd vfs.FileDescription
vfs.FileDescriptionDefaultImpl
vfs.DentryMetadataFileDescriptionImpl
vfs.LockFD
socket.SendReceiveTimeout
*waiter.Queue
family int
Endpoint tcpip.Endpoint
skType linux.SockType
protocol int
namespace *inet.Namespace
mu sync.Mutex `state:"nosave"`
// readWriter is an optimization to avoid allocations.
// +checklocks:mu
readWriter usermem.IOSequenceReadWriter `state:"nosave"`
// readMu protects access to the below fields.
readMu sync.Mutex `state:"nosave"`
// sockOptTimestamp corresponds to SO_TIMESTAMP. When true, timestamps
// of returned messages can be returned via control messages. When
// false, the same timestamp is instead stored and can be read via the
// SIOCGSTAMP ioctl. It is protected by readMu. See socket(7).
sockOptTimestamp bool
// timestampValid indicates whether timestamp for SIOCGSTAMP has been
// set. It is protected by readMu.
timestampValid bool
// timestamp holds the timestamp to use with SIOCTSTAMP. It is only
// valid when timestampValid is true. It is protected by readMu.
timestamp time.Time `state:".(int64)"`
// TODO(b/153685824): Move this to SocketOptions.
// sockOptInq corresponds to TCP_INQ.
sockOptInq bool
}
var _ = socket.Socket(&sock{})
// New creates a new endpoint socket.
func New(t *kernel.Task, family int, skType linux.SockType, protocol int, queue *waiter.Queue, endpoint tcpip.Endpoint) (*vfs.FileDescription, *syserr.Error) {
if skType == linux.SOCK_STREAM {
endpoint.SocketOptions().SetDelayOption(true)
}
mnt := t.Kernel().SocketMount()
d := sockfs.NewDentry(t, mnt)
defer d.DecRef(t)
namespace := t.NetworkNamespace()
s := &sock{
Queue: queue,
family: family,
Endpoint: endpoint,
skType: skType,
protocol: protocol,
namespace: namespace,
}
s.LockFD.Init(&vfs.FileLocks{})
vfsfd := &s.vfsfd
if err := vfsfd.Init(s, linux.O_RDWR, mnt, d, &vfs.FileDescriptionOptions{
DenyPRead: true,
DenyPWrite: true,
UseDentryMetadata: true,
}); err != nil {
return nil, syserr.FromError(err)
}
namespace.IncRef()
return vfsfd, nil
}
// Release implements vfs.FileDescriptionImpl.Release.
func (s *sock) Release(ctx context.Context) {
kernel.KernelFromContext(ctx).DeleteSocket(&s.vfsfd)
e, ch := waiter.NewChannelEntry(waiter.EventHUp | waiter.EventErr)
s.EventRegister(&e)
defer s.EventUnregister(&e)
s.Endpoint.Close()
// SO_LINGER option is valid only for TCP. For other socket types
// return after endpoint close.
if family, skType, _ := s.Type(); skType == linux.SOCK_STREAM && (family == linux.AF_INET || family == linux.AF_INET6) {
v := s.Endpoint.SocketOptions().GetLinger()
// The case for zero timeout is handled in tcp endpoint close function.
// Close is blocked until either:
// 1. The endpoint state is not in any of the states: FIN-WAIT1,
// CLOSING and LAST_ACK.
// 2. Timeout is reached.
if v.Enabled && v.Timeout != 0 {
t := kernel.TaskFromContext(ctx)
start := t.Kernel().MonotonicClock().Now()
deadline := start.Add(v.Timeout)
_ = t.BlockWithDeadline(ch, true, deadline)
}
}
s.namespace.DecRef(ctx)
}
// Epollable implements FileDescriptionImpl.Epollable.
func (s *sock) Epollable() bool {
return true
}
// Read implements vfs.FileDescriptionImpl.
func (s *sock) Read(ctx context.Context, dst usermem.IOSequence, opts vfs.ReadOptions) (int64, error) {
// All flags other than RWF_NOWAIT should be ignored.
// TODO(gvisor.dev/issue/2601): Support RWF_NOWAIT.
if opts.Flags != 0 {
return 0, linuxerr.EOPNOTSUPP
}
if dst.NumBytes() == 0 {
return 0, nil
}
n, _, _, _, _, err := s.nonBlockingRead(ctx, dst, false, false, false)
if err == syserr.ErrWouldBlock {
return int64(n), linuxerr.ErrWouldBlock
}
if err != nil {
return 0, err.ToError()
}
return int64(n), nil
}
// Write implements vfs.FileDescriptionImpl.
func (s *sock) Write(ctx context.Context, src usermem.IOSequence, opts vfs.WriteOptions) (int64, error) {
// All flags other than RWF_NOWAIT should be ignored.
// TODO(gvisor.dev/issue/2601): Support RWF_NOWAIT.
if opts.Flags != 0 {
return 0, linuxerr.EOPNOTSUPP
}
var n int64
var err tcpip.Error
switch s.Endpoint.(type) {
case *tcp.Endpoint:
s.mu.Lock()
s.readWriter.Init(ctx, src)
n, err = s.Endpoint.Write(&s.readWriter, tcpip.WriteOptions{})
s.mu.Unlock()
default:
n, err = s.Endpoint.Write(src.Reader(ctx), tcpip.WriteOptions{})
}
if _, ok := err.(*tcpip.ErrWouldBlock); ok {
return 0, linuxerr.ErrWouldBlock
}
if err != nil {
return 0, syserr.TranslateNetstackError(err).ToError()
}
if n < src.NumBytes() {
return n, linuxerr.ErrWouldBlock
}
return n, nil
}
// Accept implements the linux syscall accept(2) for sockets backed by
// tcpip.Endpoint.
func (s *sock) Accept(t *kernel.Task, peerRequested bool, flags int, blocking bool) (int32, linux.SockAddr, uint32, *syserr.Error) {
// Issue the accept request to get the new endpoint.
var peerAddr *tcpip.FullAddress
if peerRequested {
peerAddr = &tcpip.FullAddress{}
}
ep, wq, terr := s.Endpoint.Accept(peerAddr)
if terr != nil {
if _, ok := terr.(*tcpip.ErrWouldBlock); !ok || !blocking {
return 0, nil, 0, syserr.TranslateNetstackError(terr)
}
var err *syserr.Error
ep, wq, err = s.blockingAccept(t, peerAddr)
if err != nil {
return 0, nil, 0, err
}
}
ns, err := New(t, s.family, s.skType, s.protocol, wq, ep)
if err != nil {
return 0, nil, 0, err
}
defer ns.DecRef(t)
if err := ns.SetStatusFlags(t, t.Credentials(), uint32(flags&linux.SOCK_NONBLOCK)); err != nil {
return 0, nil, 0, syserr.FromError(err)
}
var addr linux.SockAddr
var addrLen uint32
if peerAddr != nil {
// Get address of the peer and write it to peer slice.
addr, addrLen = socket.ConvertAddress(s.family, *peerAddr)
}
fd, e := t.NewFDFrom(0, ns, kernel.FDFlags{
CloseOnExec: flags&linux.SOCK_CLOEXEC != 0,
})
t.Kernel().RecordSocket(ns)
return fd, addr, addrLen, syserr.FromError(e)
}
// GetSockOpt implements the linux syscall getsockopt(2) for sockets backed by
// tcpip.Endpoint.
func (s *sock) GetSockOpt(t *kernel.Task, level, name int, outPtr hostarch.Addr, outLen int) (marshal.Marshallable, *syserr.Error) {
// TODO(b/78348848): Unlike other socket options, SO_TIMESTAMP is
// implemented specifically for netstack.Socket rather than
// commonEndpoint. commonEndpoint should be extended to support socket
// options where the implementation is not shared, as unix sockets need
// their own support for SO_TIMESTAMP.
if level == linux.SOL_SOCKET && name == linux.SO_TIMESTAMP {
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
val := primitive.Int32(0)
s.readMu.Lock()
defer s.readMu.Unlock()
if s.sockOptTimestamp {
val = 1
}
return &val, nil
}
if level == linux.SOL_TCP && name == linux.TCP_INQ {
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
val := primitive.Int32(0)
s.readMu.Lock()
defer s.readMu.Unlock()
if s.sockOptInq {
val = 1
}
return &val, nil
}
return GetSockOpt(t, s, s.Endpoint, s.family, s.skType, level, name, outPtr, outLen)
}
// SetSockOpt implements the linux syscall setsockopt(2) for sockets backed by
// tcpip.Endpoint.
func (s *sock) SetSockOpt(t *kernel.Task, level int, name int, optVal []byte) *syserr.Error {
// TODO(b/78348848): Unlike other socket options, SO_TIMESTAMP is
// implemented specifically for netstack.Socket rather than
// commonEndpoint. commonEndpoint should be extended to support socket
// options where the implementation is not shared, as unix sockets need
// their own support for SO_TIMESTAMP.
if level == linux.SOL_SOCKET && name == linux.SO_TIMESTAMP {
if len(optVal) < sizeOfInt32 {
return syserr.ErrInvalidArgument
}
s.readMu.Lock()
defer s.readMu.Unlock()
s.sockOptTimestamp = hostarch.ByteOrder.Uint32(optVal) != 0
return nil
}
if level == linux.SOL_TCP && name == linux.TCP_INQ {
if len(optVal) < sizeOfInt32 {
return syserr.ErrInvalidArgument
}
s.readMu.Lock()
defer s.readMu.Unlock()
s.sockOptInq = hostarch.ByteOrder.Uint32(optVal) != 0
return nil
}
return SetSockOpt(t, s, s.Endpoint, level, name, optVal)
}
var sockAddrInetSize = (*linux.SockAddrInet)(nil).SizeBytes()
var sockAddrInet6Size = (*linux.SockAddrInet6)(nil).SizeBytes()
var sockAddrLinkSize = (*linux.SockAddrLink)(nil).SizeBytes()
// minSockAddrLen returns the minimum length in bytes of a socket address for
// the socket's family.
func (s *sock) minSockAddrLen() int {
const addressFamilySize = 2
switch s.family {
case linux.AF_UNIX:
return addressFamilySize
case linux.AF_INET:
return sockAddrInetSize
case linux.AF_INET6:
return sockAddrInet6Size
case linux.AF_PACKET:
return sockAddrLinkSize
case linux.AF_UNSPEC:
return addressFamilySize
default:
panic(fmt.Sprintf("s.family unrecognized = %d", s.family))
}
}
func (s *sock) isPacketBased() bool {
return s.skType == linux.SOCK_DGRAM || s.skType == linux.SOCK_SEQPACKET || s.skType == linux.SOCK_RDM || s.skType == linux.SOCK_RAW
}
// Readiness returns a mask of ready events for socket s.
func (s *sock) Readiness(mask waiter.EventMask) waiter.EventMask {
return s.Endpoint.Readiness(mask)
}
// checkFamily returns true iff the specified address family may be used with
// the socket.
//
// If exact is true, then the specified address family must be an exact match
// with the socket's family.
func (s *sock) checkFamily(family uint16, exact bool) bool {
if family == uint16(s.family) {
return true
}
if !exact && family == linux.AF_INET && s.family == linux.AF_INET6 {
if !s.Endpoint.SocketOptions().GetV6Only() {
return true
}
}
return false
}
// mapFamily maps the AF_INET ANY address to the IPv4-mapped IPv6 ANY if the
// receiver's family is AF_INET6.
//
// This is a hack to work around the fact that both IPv4 and IPv6 ANY are
// represented by the empty string.
//
// TODO(gvisor.dev/issue/1556): remove this function.
func (s *sock) mapFamily(addr tcpip.FullAddress, family uint16) tcpip.FullAddress {
if addr.Addr.BitLen() == 0 && s.family == linux.AF_INET6 && family == linux.AF_INET {
addr.Addr = tcpip.AddrFrom16([16]byte{0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0x00, 0x00, 0x00, 0x00})
}
return addr
}
// Connect implements the linux syscall connect(2) for sockets backed by
// tpcip.Endpoint.
func (s *sock) Connect(t *kernel.Task, sockaddr []byte, blocking bool) *syserr.Error {
addr, family, err := socket.AddressAndFamily(sockaddr)
if err != nil {
return err
}
if family == linux.AF_UNSPEC {
err := s.Endpoint.Disconnect()
if _, ok := err.(*tcpip.ErrNotSupported); ok {
return syserr.ErrAddressFamilyNotSupported
}
return syserr.TranslateNetstackError(err)
}
if !s.checkFamily(family, false /* exact */) {
return syserr.ErrInvalidArgument
}
addr = s.mapFamily(addr, family)
// Always return right away in the non-blocking case.
if !blocking {
return syserr.TranslateNetstackError(s.Endpoint.Connect(addr))
}
// Register for notification when the endpoint becomes writable, then
// initiate the connection.
e, ch := waiter.NewChannelEntry(waiter.WritableEvents)
s.EventRegister(&e)
defer s.EventUnregister(&e)
switch err := s.Endpoint.Connect(addr); err.(type) {
case *tcpip.ErrConnectStarted, *tcpip.ErrAlreadyConnecting:
case *tcpip.ErrNoPortAvailable:
if (s.family == unix.AF_INET || s.family == unix.AF_INET6) && s.skType == linux.SOCK_STREAM {
// TCP unlike UDP returns EADDRNOTAVAIL when it can't
// find an available local ephemeral port.
return syserr.ErrAddressNotAvailable
}
return syserr.TranslateNetstackError(err)
default:
return syserr.TranslateNetstackError(err)
}
// It's pending, so we have to wait for a notification, and fetch the
// result once the wait completes.
if err := t.Block(ch); err != nil {
return syserr.FromError(err)
}
// Call Connect() again after blocking to find connect's result.
return syserr.TranslateNetstackError(s.Endpoint.Connect(addr))
}
// Bind implements the linux syscall bind(2) for sockets backed by
// tcpip.Endpoint.
func (s *sock) Bind(_ *kernel.Task, sockaddr []byte) *syserr.Error {
if len(sockaddr) < 2 {
return syserr.ErrInvalidArgument
}
family := hostarch.ByteOrder.Uint16(sockaddr)
var addr tcpip.FullAddress
// Bind for AF_PACKET requires only family, protocol and ifindex.
// In function AddressAndFamily, we check the address length which is
// not needed for AF_PACKET bind.
if family == linux.AF_PACKET {
var a linux.SockAddrLink
if len(sockaddr) < sockAddrLinkSize {
return syserr.ErrInvalidArgument
}
a.UnmarshalBytes(sockaddr)
addr = tcpip.FullAddress{
NIC: tcpip.NICID(a.InterfaceIndex),
Addr: tcpip.AddrFrom16Slice(append(
a.HardwareAddr[:header.EthernetAddressSize],
[]byte{0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}...,
)),
Port: socket.Ntohs(a.Protocol),
}
} else {
if s.minSockAddrLen() > len(sockaddr) {
return syserr.ErrInvalidArgument
}
var err *syserr.Error
addr, family, err = socket.AddressAndFamily(sockaddr)
if err != nil {
return err
}
if !s.checkFamily(family, true /* exact */) {
return syserr.ErrAddressFamilyNotSupported
}
addr = s.mapFamily(addr, family)
}
// Issue the bind request to the endpoint.
err := s.Endpoint.Bind(addr)
if _, ok := err.(*tcpip.ErrNoPortAvailable); ok {
// Bind always returns EADDRINUSE irrespective of if the specified port was
// already bound or if an ephemeral port was requested but none were
// available.
//
// *tcpip.ErrNoPortAvailable is mapped to EAGAIN in syserr package because
// UDP connect returns EAGAIN on ephemeral port exhaustion.
//
// TCP connect returns EADDRNOTAVAIL on ephemeral port exhaustion.
err = &tcpip.ErrPortInUse{}
}
return syserr.TranslateNetstackError(err)
}
// Listen implements the linux syscall listen(2) for sockets backed by
// tcpip.Endpoint.
func (s *sock) Listen(_ *kernel.Task, backlog int) *syserr.Error {
if err := s.Endpoint.Listen(backlog); err != nil {
return syserr.TranslateNetstackError(err)
}
if !socket.IsTCP(s) {
return nil
}
// Emit SentryTCPListenEvent with the bound port for tcp sockets.
addr, err := s.Endpoint.GetLocalAddress()
if err != nil {
panic(fmt.Sprintf("GetLocalAddress failed for tcp socket: %s", err))
}
eventchannel.Emit(&epb.SentryTcpListenEvent{
Port: proto.Int32(int32(addr.Port)),
})
return nil
}
// blockingAccept implements a blocking version of accept(2), that is, if no
// connections are ready to be accept, it will block until one becomes ready.
func (s *sock) blockingAccept(t *kernel.Task, peerAddr *tcpip.FullAddress) (tcpip.Endpoint, *waiter.Queue, *syserr.Error) {
// Register for notifications.
e, ch := waiter.NewChannelEntry(waiter.ReadableEvents)
s.EventRegister(&e)
defer s.EventUnregister(&e)
// Try to accept the connection again; if it fails, then wait until we
// get a notification.
for {
ep, wq, err := s.Endpoint.Accept(peerAddr)
if _, ok := err.(*tcpip.ErrWouldBlock); !ok {
return ep, wq, syserr.TranslateNetstackError(err)
}
if err := t.Block(ch); err != nil {
return nil, nil, syserr.FromError(err)
}
}
}
// ConvertShutdown converts Linux shutdown flags into tcpip shutdown flags.
func ConvertShutdown(how int) (tcpip.ShutdownFlags, *syserr.Error) {
var f tcpip.ShutdownFlags
switch how {
case linux.SHUT_RD:
f = tcpip.ShutdownRead
case linux.SHUT_WR:
f = tcpip.ShutdownWrite
case linux.SHUT_RDWR:
f = tcpip.ShutdownRead | tcpip.ShutdownWrite
default:
return 0, syserr.ErrInvalidArgument
}
return f, nil
}
// Shutdown implements the linux syscall shutdown(2) for sockets backed by
// tcpip.Endpoint.
func (s *sock) Shutdown(_ *kernel.Task, how int) *syserr.Error {
f, err := ConvertShutdown(how)
if err != nil {
return err
}
// Issue shutdown request.
return syserr.TranslateNetstackError(s.Endpoint.Shutdown(f))
}
// GetSockOpt can be used to implement the linux syscall getsockopt(2) for
// sockets backed by a commonEndpoint.
func GetSockOpt(t *kernel.Task, s socket.Socket, ep commonEndpoint, family int, skType linux.SockType, level, name int, outPtr hostarch.Addr, outLen int) (marshal.Marshallable, *syserr.Error) {
switch level {
case linux.SOL_SOCKET:
return getSockOptSocket(t, s, ep, family, skType, name, outLen)
case linux.SOL_TCP:
return getSockOptTCP(t, s, ep, name, outLen)
case linux.SOL_IPV6:
return getSockOptIPv6(t, s, ep, name, outPtr, outLen)
case linux.SOL_IP:
return getSockOptIP(t, s, ep, name, outPtr, outLen, family)
case linux.SOL_ICMPV6:
return getSockOptICMPv6(t, s, ep, name, outLen)
case linux.SOL_UDP,
linux.SOL_RAW,
linux.SOL_PACKET:
// Not supported.
}
return nil, syserr.ErrProtocolNotAvailable
}
func boolToInt32(v bool) int32 {
if v {
return 1
}
return 0
}
// getSockOptSocket implements GetSockOpt when level is SOL_SOCKET.
func getSockOptSocket(t *kernel.Task, s socket.Socket, ep commonEndpoint, family int, _ linux.SockType, name, outLen int) (marshal.Marshallable, *syserr.Error) {
// TODO(b/124056281): Stop rejecting short optLen values in getsockopt.
switch name {
case linux.SO_ERROR:
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
// Get the last error and convert it.
err := ep.SocketOptions().GetLastError()
if err == nil {
optP := primitive.Int32(0)
return &optP, nil
}
optP := primitive.Int32(syserr.TranslateNetstackError(err).ToLinux())
return &optP, nil
case linux.SO_PEERCRED:
if family != linux.AF_UNIX || outLen < unix.SizeofUcred {
return nil, syserr.ErrInvalidArgument
}
tcred := t.Credentials()
creds := linux.ControlMessageCredentials{
PID: int32(t.ThreadGroup().ID()),
UID: uint32(tcred.EffectiveKUID.In(tcred.UserNamespace).OrOverflow()),
GID: uint32(tcred.EffectiveKGID.In(tcred.UserNamespace).OrOverflow()),
}
return &creds, nil
case linux.SO_PASSCRED:
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
v := primitive.Int32(boolToInt32(ep.SocketOptions().GetPassCred()))
return &v, nil
case linux.SO_SNDBUF:
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
size := ep.SocketOptions().GetSendBufferSize()
if size > math.MaxInt32 {
size = math.MaxInt32
}
sizeP := primitive.Int32(size)
return &sizeP, nil
case linux.SO_RCVBUF:
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
size := ep.SocketOptions().GetReceiveBufferSize()
if size > math.MaxInt32 {
size = math.MaxInt32
}
sizeP := primitive.Int32(size)
return &sizeP, nil
case linux.SO_REUSEADDR:
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
v := primitive.Int32(boolToInt32(ep.SocketOptions().GetReuseAddress()))
return &v, nil
case linux.SO_REUSEPORT:
if outLen < sizeOfInt32 {
return nil, syserr.ErrInvalidArgument
}
v := primitive.Int32(boolToInt32(ep.SocketOptions().GetReusePort()))
return &v, nil