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
|
// 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 stack
import (
"fmt"
"inet.af/netstack/sync"
"inet.af/netstack/tcpip"
"inet.af/netstack/tcpip/hash/jenkins"
"inet.af/netstack/tcpip/header"
"inet.af/netstack/tcpip/ports"
)
type protocolIDs struct {
network tcpip.NetworkProtocolNumber
transport tcpip.TransportProtocolNumber
}
// transportEndpoints manages all endpoints of a given protocol. It has its own
// mutex so as to reduce interference between protocols.
type transportEndpoints struct {
mu sync.RWMutex
// +checklocks:mu
endpoints map[TransportEndpointID]*endpointsByNIC
// rawEndpoints contains endpoints for raw sockets, which receive all
// traffic of a given protocol regardless of port.
//
// +checklocks:mu
rawEndpoints []RawTransportEndpoint
}
// unregisterEndpoint unregisters the endpoint with the given id such that it
// won't receive any more packets.
func (eps *transportEndpoints) unregisterEndpoint(id TransportEndpointID, ep TransportEndpoint, flags ports.Flags, bindToDevice tcpip.NICID) {
eps.mu.Lock()
defer eps.mu.Unlock()
epsByNIC, ok := eps.endpoints[id]
if !ok {
return
}
if !epsByNIC.unregisterEndpoint(bindToDevice, ep, flags) {
return
}
delete(eps.endpoints, id)
}
func (eps *transportEndpoints) transportEndpoints() []TransportEndpoint {
eps.mu.RLock()
defer eps.mu.RUnlock()
es := make([]TransportEndpoint, 0, len(eps.endpoints))
for _, e := range eps.endpoints {
es = append(es, e.transportEndpoints()...)
}
return es
}
// iterEndpointsLocked yields all endpointsByNIC in eps that match id, in
// descending order of match quality. If a call to yield returns false,
// iterEndpointsLocked stops iteration and returns immediately.
//
// +checklocksread:eps.mu
func (eps *transportEndpoints) iterEndpointsLocked(id TransportEndpointID, yield func(*endpointsByNIC) bool) {
// Try to find a match with the id as provided.
if ep, ok := eps.endpoints[id]; ok {
if !yield(ep) {
return
}
}
// Try to find a match with the id minus the local address.
nid := id
nid.LocalAddress = ""
if ep, ok := eps.endpoints[nid]; ok {
if !yield(ep) {
return
}
}
// Try to find a match with the id minus the remote part.
nid.LocalAddress = id.LocalAddress
nid.RemoteAddress = ""
nid.RemotePort = 0
if ep, ok := eps.endpoints[nid]; ok {
if !yield(ep) {
return
}
}
// Try to find a match with only the local port.
nid.LocalAddress = ""
if ep, ok := eps.endpoints[nid]; ok {
if !yield(ep) {
return
}
}
}
// findAllEndpointsLocked returns all endpointsByNIC in eps that match id, in
// descending order of match quality.
//
// +checklocksread:eps.mu
func (eps *transportEndpoints) findAllEndpointsLocked(id TransportEndpointID) []*endpointsByNIC {
var matchedEPs []*endpointsByNIC
eps.iterEndpointsLocked(id, func(ep *endpointsByNIC) bool {
matchedEPs = append(matchedEPs, ep)
return true
})
return matchedEPs
}
// findEndpointLocked returns the endpoint that most closely matches the given id.
//
// +checklocksread:eps.mu
func (eps *transportEndpoints) findEndpointLocked(id TransportEndpointID) *endpointsByNIC {
var matchedEP *endpointsByNIC
eps.iterEndpointsLocked(id, func(ep *endpointsByNIC) bool {
matchedEP = ep
return false
})
return matchedEP
}
type endpointsByNIC struct {
// seed is a random secret for a jenkins hash.
seed uint32
mu sync.RWMutex
// +checklocks:mu
endpoints map[tcpip.NICID]*multiPortEndpoint
}
func (epsByNIC *endpointsByNIC) transportEndpoints() []TransportEndpoint {
epsByNIC.mu.RLock()
defer epsByNIC.mu.RUnlock()
var eps []TransportEndpoint
for _, ep := range epsByNIC.endpoints {
eps = append(eps, ep.transportEndpoints()...)
}
return eps
}
// handlePacket is called by the stack when new packets arrive to this transport
// endpoint. It returns false if the packet could not be matched to any
// transport endpoint, true otherwise.
func (epsByNIC *endpointsByNIC) handlePacket(id TransportEndpointID, pkt *PacketBuffer) bool {
epsByNIC.mu.RLock()
mpep, ok := epsByNIC.endpoints[pkt.NICID]
if !ok {
if mpep, ok = epsByNIC.endpoints[0]; !ok {
epsByNIC.mu.RUnlock() // Don't use defer for performance reasons.
return false
}
}
// If this is a broadcast or multicast datagram, deliver the datagram to all
// endpoints bound to the right device.
if isInboundMulticastOrBroadcast(pkt, id.LocalAddress) {
mpep.handlePacketAll(id, pkt)
epsByNIC.mu.RUnlock() // Don't use defer for performance reasons.
return true
}
// multiPortEndpoints are guaranteed to have at least one element.
transEP := mpep.selectEndpoint(id, epsByNIC.seed)
if queuedProtocol, mustQueue := mpep.demux.queuedProtocols[protocolIDs{mpep.netProto, mpep.transProto}]; mustQueue {
queuedProtocol.QueuePacket(transEP, id, pkt)
epsByNIC.mu.RUnlock()
return true
}
transEP.HandlePacket(id, pkt)
epsByNIC.mu.RUnlock() // Don't use defer for performance reasons.
return true
}
// handleError delivers an error to the transport endpoint identified by id.
func (epsByNIC *endpointsByNIC) handleError(n *nic, id TransportEndpointID, transErr TransportError, pkt *PacketBuffer) {
epsByNIC.mu.RLock()
defer epsByNIC.mu.RUnlock()
mpep, ok := epsByNIC.endpoints[n.ID()]
if !ok {
mpep, ok = epsByNIC.endpoints[0]
}
if !ok {
return
}
// TODO(eyalsoha): Why don't we look at id to see if this packet needs to
// broadcast like we are doing with handlePacket above?
// multiPortEndpoints are guaranteed to have at least one element.
mpep.selectEndpoint(id, epsByNIC.seed).HandleError(transErr, pkt)
}
// registerEndpoint returns true if it succeeds. It fails and returns
// false if ep already has an element with the same key.
func (epsByNIC *endpointsByNIC) registerEndpoint(d *transportDemuxer, netProto tcpip.NetworkProtocolNumber, transProto tcpip.TransportProtocolNumber, t TransportEndpoint, flags ports.Flags, bindToDevice tcpip.NICID) tcpip.Error {
epsByNIC.mu.Lock()
defer epsByNIC.mu.Unlock()
multiPortEp, ok := epsByNIC.endpoints[bindToDevice]
if !ok {
multiPortEp = &multiPortEndpoint{
demux: d,
netProto: netProto,
transProto: transProto,
}
}
if err := multiPortEp.singleRegisterEndpoint(t, flags); err != nil {
return err
}
// Only add this newly created multiportEndpoint if the singleRegisterEndpoint
// succeeded.
if !ok {
epsByNIC.endpoints[bindToDevice] = multiPortEp
}
return nil
}
func (epsByNIC *endpointsByNIC) checkEndpoint(flags ports.Flags, bindToDevice tcpip.NICID) tcpip.Error {
epsByNIC.mu.RLock()
defer epsByNIC.mu.RUnlock()
multiPortEp, ok := epsByNIC.endpoints[bindToDevice]
if !ok {
return nil
}
return multiPortEp.singleCheckEndpoint(flags)
}
// unregisterEndpoint returns true if endpointsByNIC has to be unregistered.
func (epsByNIC *endpointsByNIC) unregisterEndpoint(bindToDevice tcpip.NICID, t TransportEndpoint, flags ports.Flags) bool {
epsByNIC.mu.Lock()
defer epsByNIC.mu.Unlock()
multiPortEp, ok := epsByNIC.endpoints[bindToDevice]
if !ok {
return false
}
if multiPortEp.unregisterEndpoint(t, flags) {
delete(epsByNIC.endpoints, bindToDevice)
}
return len(epsByNIC.endpoints) == 0
}
// transportDemuxer demultiplexes packets targeted at a transport endpoint
// (i.e., after they've been parsed by the network layer). It does two levels
// of demultiplexing: first based on the network and transport protocols, then
// based on endpoints IDs. It should only be instantiated via
// newTransportDemuxer.
type transportDemuxer struct {
stack *Stack
// protocol is immutable.
protocol map[protocolIDs]*transportEndpoints
queuedProtocols map[protocolIDs]queuedTransportProtocol
}
// queuedTransportProtocol if supported by a protocol implementation will cause
// the dispatcher to delivery packets to the QueuePacket method instead of
// calling HandlePacket directly on the endpoint.
type queuedTransportProtocol interface {
QueuePacket(ep TransportEndpoint, id TransportEndpointID, pkt *PacketBuffer)
}
func newTransportDemuxer(stack *Stack) *transportDemuxer {
d := &transportDemuxer{
stack: stack,
protocol: make(map[protocolIDs]*transportEndpoints),
queuedProtocols: make(map[protocolIDs]queuedTransportProtocol),
}
// Add each network and transport pair to the demuxer.
for netProto := range stack.networkProtocols {
for proto := range stack.transportProtocols {
protoIDs := protocolIDs{netProto, proto}
d.protocol[protoIDs] = &transportEndpoints{
endpoints: make(map[TransportEndpointID]*endpointsByNIC),
}
qTransProto, isQueued := (stack.transportProtocols[proto].proto).(queuedTransportProtocol)
if isQueued {
d.queuedProtocols[protoIDs] = qTransProto
}
}
}
return d
}
// registerEndpoint registers the given endpoint with the dispatcher such that
// packets that match the endpoint ID are delivered to it.
func (d *transportDemuxer) registerEndpoint(netProtos []tcpip.NetworkProtocolNumber, protocol tcpip.TransportProtocolNumber, id TransportEndpointID, ep TransportEndpoint, flags ports.Flags, bindToDevice tcpip.NICID) tcpip.Error {
for i, n := range netProtos {
if err := d.singleRegisterEndpoint(n, protocol, id, ep, flags, bindToDevice); err != nil {
d.unregisterEndpoint(netProtos[:i], protocol, id, ep, flags, bindToDevice)
return err
}
}
return nil
}
// checkEndpoint checks if an endpoint can be registered with the dispatcher.
func (d *transportDemuxer) checkEndpoint(netProtos []tcpip.NetworkProtocolNumber, protocol tcpip.TransportProtocolNumber, id TransportEndpointID, flags ports.Flags, bindToDevice tcpip.NICID) tcpip.Error {
for _, n := range netProtos {
if err := d.singleCheckEndpoint(n, protocol, id, flags, bindToDevice); err != nil {
return err
}
}
return nil
}
// multiPortEndpoint is a container for TransportEndpoints which are bound to
// the same pair of address and port. endpointsArr always has at least one
// element.
//
// FIXME(gvisor.dev/issue/873): Restore this properly. Currently, we just save
// this to ensure that the underlying endpoints get saved/restored, but not not
// use the restored copy.
//
// +stateify savable
type multiPortEndpoint struct {
demux *transportDemuxer
netProto tcpip.NetworkProtocolNumber
transProto tcpip.TransportProtocolNumber
flags ports.FlagCounter
mu sync.RWMutex `state:"nosave"`
// endpoints stores the transport endpoints in the order in which they
// were bound. This is required for UDP SO_REUSEADDR.
//
// +checklocks:mu
endpoints []TransportEndpoint
}
func (ep *multiPortEndpoint) transportEndpoints() []TransportEndpoint {
ep.mu.RLock()
eps := append([]TransportEndpoint(nil), ep.endpoints...)
ep.mu.RUnlock()
return eps
}
// reciprocalScale scales a value into range [0, n).
//
// This is similar to val % n, but faster.
// See http://lemire.me/blog/2016/06/27/a-fast-alternative-to-the-modulo-reduction/
func reciprocalScale(val, n uint32) uint32 {
return uint32((uint64(val) * uint64(n)) >> 32)
}
// selectEndpoint calculates a hash of destination and source addresses and
// ports then uses it to select a socket. In this case, all packets from one
// address will be sent to same endpoint.
func (ep *multiPortEndpoint) selectEndpoint(id TransportEndpointID, seed uint32) TransportEndpoint {
ep.mu.RLock()
defer ep.mu.RUnlock()
if len(ep.endpoints) == 1 {
return ep.endpoints[0]
}
if ep.flags.SharedFlags().ToFlags().Effective().MostRecent {
return ep.endpoints[len(ep.endpoints)-1]
}
payload := []byte{
byte(id.LocalPort),
byte(id.LocalPort >> 8),
byte(id.RemotePort),
byte(id.RemotePort >> 8),
}
h := jenkins.Sum32(seed)
h.Write(payload)
h.Write([]byte(id.LocalAddress))
h.Write([]byte(id.RemoteAddress))
hash := h.Sum32()
idx := reciprocalScale(hash, uint32(len(ep.endpoints)))
return ep.endpoints[idx]
}
func (ep *multiPortEndpoint) handlePacketAll(id TransportEndpointID, pkt *PacketBuffer) {
ep.mu.RLock()
queuedProtocol, mustQueue := ep.demux.queuedProtocols[protocolIDs{ep.netProto, ep.transProto}]
// HandlePacket may modify pkt, so each endpoint needs
// its own copy except for the final one.
for _, endpoint := range ep.endpoints[:len(ep.endpoints)-1] {
clone := pkt.Clone()
if mustQueue {
queuedProtocol.QueuePacket(endpoint, id, clone)
} else {
endpoint.HandlePacket(id, clone)
}
clone.DecRef()
}
if endpoint := ep.endpoints[len(ep.endpoints)-1]; mustQueue {
queuedProtocol.QueuePacket(endpoint, id, pkt)
} else {
endpoint.HandlePacket(id, pkt)
}
ep.mu.RUnlock() // Don't use defer for performance reasons.
}
// singleRegisterEndpoint tries to add an endpoint to the multiPortEndpoint
// list. The list might be empty already.
func (ep *multiPortEndpoint) singleRegisterEndpoint(t TransportEndpoint, flags ports.Flags) tcpip.Error {
ep.mu.Lock()
defer ep.mu.Unlock()
bits := flags.Bits() & ports.MultiBindFlagMask
if len(ep.endpoints) != 0 {
// If it was previously bound, we need to check if we can bind again.
if ep.flags.TotalRefs() > 0 && bits&ep.flags.SharedFlags() == 0 {
return &tcpip.ErrPortInUse{}
}
}
ep.endpoints = append(ep.endpoints, t)
ep.flags.AddRef(bits)
return nil
}
func (ep *multiPortEndpoint) singleCheckEndpoint(flags ports.Flags) tcpip.Error {
ep.mu.RLock()
defer ep.mu.RUnlock()
bits := flags.Bits() & ports.MultiBindFlagMask
if len(ep.endpoints) != 0 {
// If it was previously bound, we need to check if we can bind again.
if ep.flags.TotalRefs() > 0 && bits&ep.flags.SharedFlags() == 0 {
return &tcpip.ErrPortInUse{}
}
}
return nil
}
// unregisterEndpoint returns true if multiPortEndpoint has to be unregistered.
func (ep *multiPortEndpoint) unregisterEndpoint(t TransportEndpoint, flags ports.Flags) bool {
ep.mu.Lock()
defer ep.mu.Unlock()
for i, endpoint := range ep.endpoints {
if endpoint == t {
copy(ep.endpoints[i:], ep.endpoints[i+1:])
ep.endpoints[len(ep.endpoints)-1] = nil
ep.endpoints = ep.endpoints[:len(ep.endpoints)-1]
ep.flags.DropRef(flags.Bits() & ports.MultiBindFlagMask)
break
}
}
return len(ep.endpoints) == 0
}
func (d *transportDemuxer) singleRegisterEndpoint(netProto tcpip.NetworkProtocolNumber, protocol tcpip.TransportProtocolNumber, id TransportEndpointID, ep TransportEndpoint, flags ports.Flags, bindToDevice tcpip.NICID) tcpip.Error {
if id.RemotePort != 0 {
// SO_REUSEPORT only applies to bound/listening endpoints.
flags.LoadBalanced = false
}
eps, ok := d.protocol[protocolIDs{netProto, protocol}]
if !ok {
return &tcpip.ErrUnknownProtocol{}
}
eps.mu.Lock()
defer eps.mu.Unlock()
epsByNIC, ok := eps.endpoints[id]
if !ok {
epsByNIC = &endpointsByNIC{
endpoints: make(map[tcpip.NICID]*multiPortEndpoint),
seed: d.stack.seed,
}
}
if err := epsByNIC.registerEndpoint(d, netProto, protocol, ep, flags, bindToDevice); err != nil {
return err
}
// Only add this newly created epsByNIC if registerEndpoint succeeded.
if !ok {
eps.endpoints[id] = epsByNIC
}
return nil
}
func (d *transportDemuxer) singleCheckEndpoint(netProto tcpip.NetworkProtocolNumber, protocol tcpip.TransportProtocolNumber, id TransportEndpointID, flags ports.Flags, bindToDevice tcpip.NICID) tcpip.Error {
if id.RemotePort != 0 {
// SO_REUSEPORT only applies to bound/listening endpoints.
flags.LoadBalanced = false
}
eps, ok := d.protocol[protocolIDs{netProto, protocol}]
if !ok {
return &tcpip.ErrUnknownProtocol{}
}
eps.mu.RLock()
defer eps.mu.RUnlock()
epsByNIC, ok := eps.endpoints[id]
if !ok {
return nil
}
return epsByNIC.checkEndpoint(flags, bindToDevice)
}
// unregisterEndpoint unregisters the endpoint with the given id such that it
// won't receive any more packets.
func (d *transportDemuxer) unregisterEndpoint(netProtos []tcpip.NetworkProtocolNumber, protocol tcpip.TransportProtocolNumber, id TransportEndpointID, ep TransportEndpoint, flags ports.Flags, bindToDevice tcpip.NICID) {
if id.RemotePort != 0 {
// SO_REUSEPORT only applies to bound/listening endpoints.
flags.LoadBalanced = false
}
for _, n := range netProtos {
if eps, ok := d.protocol[protocolIDs{n, protocol}]; ok {
eps.unregisterEndpoint(id, ep, flags, bindToDevice)
}
}
}
// deliverPacket attempts to find one or more matching transport endpoints, and
// then, if matches are found, delivers the packet to them. Returns true if
// the packet no longer needs to be handled.
func (d *transportDemuxer) deliverPacket(protocol tcpip.TransportProtocolNumber, pkt *PacketBuffer, id TransportEndpointID) bool {
eps, ok := d.protocol[protocolIDs{pkt.NetworkProtocolNumber, protocol}]
if !ok {
return false
}
// If the packet is a UDP broadcast or multicast, then find all matching
// transport endpoints.
if protocol == header.UDPProtocolNumber && isInboundMulticastOrBroadcast(pkt, id.LocalAddress) {
eps.mu.RLock()
destEPs := eps.findAllEndpointsLocked(id)
eps.mu.RUnlock()
// Fail if we didn't find at least one matching transport endpoint.
if len(destEPs) == 0 {
d.stack.stats.UDP.UnknownPortErrors.Increment()
return false
}
// handlePacket takes may modify pkt, so each endpoint needs its own
// copy except for the final one.
for _, ep := range destEPs[:len(destEPs)-1] {
clone := pkt.Clone()
ep.handlePacket(id, clone)
clone.DecRef()
}
destEPs[len(destEPs)-1].handlePacket(id, pkt)
return true
}
// If the packet is a TCP packet with a unspecified source or non-unicast
// destination address, then do nothing further and instruct the caller to do
// the same. The network layer handles address validation for specified source
// addresses.
if protocol == header.TCPProtocolNumber && (!isSpecified(id.LocalAddress) || !isSpecified(id.RemoteAddress) || isInboundMulticastOrBroadcast(pkt, id.LocalAddress)) {
// TCP can only be used to communicate between a single source and a
// single destination; the addresses must be unicast.e
d.stack.stats.TCP.InvalidSegmentsReceived.Increment()
return true
}
eps.mu.RLock()
ep := eps.findEndpointLocked(id)
eps.mu.RUnlock()
if ep == nil {
if protocol == header.UDPProtocolNumber {
d.stack.stats.UDP.UnknownPortErrors.Increment()
}
return false
}
return ep.handlePacket(id, pkt)
}
// deliverRawPacket attempts to deliver the given packet and returns whether it
// was delivered successfully.
func (d *transportDemuxer) deliverRawPacket(protocol tcpip.TransportProtocolNumber, pkt *PacketBuffer) bool {
eps, ok := d.protocol[protocolIDs{pkt.NetworkProtocolNumber, protocol}]
if !ok {
return false
}
// As in net/ipv4/ip_input.c:ip_local_deliver, attempt to deliver via
// raw endpoint first. If there are multiple raw endpoints, they all
// receive the packet.
eps.mu.RLock()
// Copy the list of raw endpoints to avoid packet handling under lock.
var rawEPs []RawTransportEndpoint
if n := len(eps.rawEndpoints); n != 0 {
rawEPs = make([]RawTransportEndpoint, n)
if m := copy(rawEPs, eps.rawEndpoints); m != n {
panic(fmt.Sprintf("unexpected copy = %d, want %d", m, n))
}
}
eps.mu.RUnlock()
for _, rawEP := range rawEPs {
// Each endpoint gets its own copy of the packet for the sake
// of save/restore.
clone := pkt.Clone()
rawEP.HandlePacket(clone)
clone.DecRef()
}
return len(rawEPs) != 0
}
// deliverError attempts to deliver the given error to the appropriate transport
// endpoint.
//
// Returns true if the error was delivered.
func (d *transportDemuxer) deliverError(n *nic, net tcpip.NetworkProtocolNumber, trans tcpip.TransportProtocolNumber, transErr TransportError, pkt *PacketBuffer, id TransportEndpointID) bool {
eps, ok := d.protocol[protocolIDs{net, trans}]
if !ok {
return false
}
eps.mu.RLock()
ep := eps.findEndpointLocked(id)
eps.mu.RUnlock()
if ep == nil {
return false
}
ep.handleError(n, id, transErr, pkt)
return true
}
// findTransportEndpoint find a single endpoint that most closely matches the provided id.
func (d *transportDemuxer) findTransportEndpoint(netProto tcpip.NetworkProtocolNumber, transProto tcpip.TransportProtocolNumber, id TransportEndpointID, nicID tcpip.NICID) TransportEndpoint {
eps, ok := d.protocol[protocolIDs{netProto, transProto}]
if !ok {
return nil
}
eps.mu.RLock()
epsByNIC := eps.findEndpointLocked(id)
if epsByNIC == nil {
eps.mu.RUnlock()
return nil
}
epsByNIC.mu.RLock()
eps.mu.RUnlock()
mpep, ok := epsByNIC.endpoints[nicID]
if !ok {
if mpep, ok = epsByNIC.endpoints[0]; !ok {
epsByNIC.mu.RUnlock() // Don't use defer for performance reasons.
return nil
}
}
ep := mpep.selectEndpoint(id, epsByNIC.seed)
epsByNIC.mu.RUnlock()
return ep
}
// registerRawEndpoint registers the given endpoint with the dispatcher such
// that packets of the appropriate protocol are delivered to it. A single
// packet can be sent to one or more raw endpoints along with a non-raw
// endpoint.
func (d *transportDemuxer) registerRawEndpoint(netProto tcpip.NetworkProtocolNumber, transProto tcpip.TransportProtocolNumber, ep RawTransportEndpoint) tcpip.Error {
eps, ok := d.protocol[protocolIDs{netProto, transProto}]
if !ok {
return &tcpip.ErrNotSupported{}
}
eps.mu.Lock()
eps.rawEndpoints = append(eps.rawEndpoints, ep)
eps.mu.Unlock()
return nil
}
// unregisterRawEndpoint unregisters the raw endpoint for the given transport
// protocol such that it won't receive any more packets.
func (d *transportDemuxer) unregisterRawEndpoint(netProto tcpip.NetworkProtocolNumber, transProto tcpip.TransportProtocolNumber, ep RawTransportEndpoint) {
eps, ok := d.protocol[protocolIDs{netProto, transProto}]
if !ok {
panic(fmt.Errorf("tried to unregister endpoint with unsupported network and transport protocol pair: %d, %d", netProto, transProto))
}
eps.mu.Lock()
for i, rawEP := range eps.rawEndpoints {
if rawEP == ep {
lastIdx := len(eps.rawEndpoints) - 1
eps.rawEndpoints[i] = eps.rawEndpoints[lastIdx]
eps.rawEndpoints[lastIdx] = nil
eps.rawEndpoints = eps.rawEndpoints[:lastIdx]
break
}
}
eps.mu.Unlock()
}
func isInboundMulticastOrBroadcast(pkt *PacketBuffer, localAddr tcpip.Address) bool {
return pkt.NetworkPacketInfo.LocalAddressBroadcast || header.IsV4MulticastAddress(localAddr) || header.IsV6MulticastAddress(localAddr)
}
func isSpecified(addr tcpip.Address) bool {
return addr != header.IPv4Any && addr != header.IPv6Any
}
|