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
|
// 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 ipv6
import (
"gvisor.dev/gvisor/pkg/tcpip"
"gvisor.dev/gvisor/pkg/tcpip/buffer"
"gvisor.dev/gvisor/pkg/tcpip/header"
"gvisor.dev/gvisor/pkg/tcpip/stack"
)
// handleControl handles the case when an ICMP packet contains the headers of
// the original packet that caused the ICMP one to be sent. This information is
// used to find out which transport endpoint must be notified about the ICMP
// packet.
func (e *endpoint) handleControl(typ stack.ControlType, extra uint32, vv buffer.VectorisedView) {
h := header.IPv6(vv.First())
// We don't use IsValid() here because ICMP only requires that up to
// 1280 bytes of the original packet be included. So it's likely that it
// is truncated, which would cause IsValid to return false.
//
// Drop packet if it doesn't have the basic IPv6 header or if the
// original source address doesn't match the endpoint's address.
if len(h) < header.IPv6MinimumSize || h.SourceAddress() != e.id.LocalAddress {
return
}
// Skip the IP header, then handle the fragmentation header if there
// is one.
vv.TrimFront(header.IPv6MinimumSize)
p := h.TransportProtocol()
if p == header.IPv6FragmentHeader {
f := header.IPv6Fragment(vv.First())
if !f.IsValid() || f.FragmentOffset() != 0 {
// We can't handle fragments that aren't at offset 0
// because they don't have the transport headers.
return
}
// Skip fragmentation header and find out the actual protocol
// number.
vv.TrimFront(header.IPv6FragmentHeaderSize)
p = f.TransportProtocol()
}
// Deliver the control packet to the transport endpoint.
e.dispatcher.DeliverTransportControlPacket(e.id.LocalAddress, h.DestinationAddress(), ProtocolNumber, p, typ, extra, vv)
}
func (e *endpoint) handleICMP(r *stack.Route, netHeader buffer.View, vv buffer.VectorisedView) {
stats := r.Stats().ICMP
sent := stats.V6PacketsSent
received := stats.V6PacketsReceived
v := vv.First()
if len(v) < header.ICMPv6MinimumSize {
received.Invalid.Increment()
return
}
h := header.ICMPv6(v)
iph := header.IPv6(netHeader)
// Validate ICMPv6 checksum before processing the packet.
//
// Only the first view in vv is accounted for by h. To account for the
// rest of vv, a shallow copy is made and the first view is removed.
// This copy is used as extra payload during the checksum calculation.
payload := vv
payload.RemoveFirst()
if got, want := h.Checksum(), header.ICMPv6Checksum(h, iph.SourceAddress(), iph.DestinationAddress(), payload); got != want {
received.Invalid.Increment()
return
}
// As per RFC 4861 sections 4.1 - 4.5, 6.1.1, 6.1.2, 7.1.1, 7.1.2 and
// 8.1, nodes MUST silently drop NDP packets where the Hop Limit field
// in the IPv6 header is not set to 255.
switch h.Type() {
case header.ICMPv6NeighborSolicit,
header.ICMPv6NeighborAdvert,
header.ICMPv6RouterSolicit,
header.ICMPv6RouterAdvert,
header.ICMPv6RedirectMsg:
if iph.HopLimit() != header.NDPHopLimit {
received.Invalid.Increment()
return
}
}
// TODO(b/112892170): Meaningfully handle all ICMP types.
switch h.Type() {
case header.ICMPv6PacketTooBig:
received.PacketTooBig.Increment()
if len(v) < header.ICMPv6PacketTooBigMinimumSize {
received.Invalid.Increment()
return
}
vv.TrimFront(header.ICMPv6PacketTooBigMinimumSize)
mtu := h.MTU()
e.handleControl(stack.ControlPacketTooBig, calculateMTU(mtu), vv)
case header.ICMPv6DstUnreachable:
received.DstUnreachable.Increment()
if len(v) < header.ICMPv6DstUnreachableMinimumSize {
received.Invalid.Increment()
return
}
vv.TrimFront(header.ICMPv6DstUnreachableMinimumSize)
switch h.Code() {
case header.ICMPv6PortUnreachable:
e.handleControl(stack.ControlPortUnreachable, 0, vv)
}
case header.ICMPv6NeighborSolicit:
received.NeighborSolicit.Increment()
if len(v) < header.ICMPv6NeighborSolicitMinimumSize {
received.Invalid.Increment()
return
}
ns := header.NDPNeighborSolicit(h.NDPPayload())
targetAddr := ns.TargetAddress()
s := r.Stack()
rxNICID := r.NICID()
isTentative, err := s.IsAddrTentative(rxNICID, targetAddr)
if err != nil {
// We will only get an error if rxNICID is unrecognized,
// which should not happen. For now short-circuit this
// packet.
//
// TODO(b/141002840): Handle this better?
return
}
if isTentative {
// If the target address is tentative and the source
// of the packet is a unicast (specified) address, then
// the source of the packet is attempting to perform
// address resolution on the target. In this case, the
// solicitation is silently ignored, as per RFC 4862
// section 5.4.3.
//
// If the target address is tentative and the source of
// the packet is the unspecified address (::), then we
// know another node is also performing DAD for the
// same address (since targetAddr is tentative for us,
// we know we are also performing DAD on it). In this
// case we let the stack know so it can handle such a
// scenario and do nothing further with the NDP NS.
if iph.SourceAddress() == header.IPv6Any {
s.DupTentativeAddrDetected(rxNICID, targetAddr)
}
// Do not handle neighbor solicitations targeted
// to an address that is tentative on the received
// NIC any further.
return
}
// At this point we know that targetAddr is not tentative on
// rxNICID so the packet is processed as defined in RFC 4861,
// as per RFC 4862 section 5.4.3.
if e.linkAddrCache.CheckLocalAddress(e.nicid, ProtocolNumber, targetAddr) == 0 {
// We don't have a useful answer; the best we can do is ignore the request.
return
}
optsSerializer := header.NDPOptionsSerializer{
header.NDPTargetLinkLayerAddressOption(r.LocalLinkAddress[:]),
}
hdr := buffer.NewPrependable(int(r.MaxHeaderLength()) + header.ICMPv6NeighborAdvertMinimumSize + int(optsSerializer.Length()))
pkt := header.ICMPv6(hdr.Prepend(header.ICMPv6NeighborAdvertSize))
pkt.SetType(header.ICMPv6NeighborAdvert)
na := header.NDPNeighborAdvert(pkt.NDPPayload())
na.SetSolicitedFlag(true)
na.SetOverrideFlag(true)
na.SetTargetAddress(targetAddr)
opts := na.Options()
opts.Serialize(optsSerializer)
// ICMPv6 Neighbor Solicit messages are always sent to
// specially crafted IPv6 multicast addresses. As a result, the
// route we end up with here has as its LocalAddress such a
// multicast address. It would be nonsense to claim that our
// source address is a multicast address, so we manually set
// the source address to the target address requested in the
// solicit message. Since that requires mutating the route, we
// must first clone it.
r := r.Clone()
defer r.Release()
r.LocalAddress = targetAddr
pkt.SetChecksum(header.ICMPv6Checksum(pkt, r.LocalAddress, r.RemoteAddress, buffer.VectorisedView{}))
// TODO(tamird/ghanan): there exists an explicit NDP option that is
// used to update the neighbor table with link addresses for a
// neighbor from an NS (see the Source Link Layer option RFC
// 4861 section 4.6.1 and section 7.2.3).
//
// Furthermore, the entirety of NDP handling here seems to be
// contradicted by RFC 4861.
e.linkAddrCache.AddLinkAddress(e.nicid, r.RemoteAddress, r.RemoteLinkAddress)
// RFC 4861 Neighbor Discovery for IP version 6 (IPv6)
//
// 7.1.2. Validation of Neighbor Advertisements
//
// The IP Hop Limit field has a value of 255, i.e., the packet
// could not possibly have been forwarded by a router.
if err := r.WritePacket(nil /* gso */, hdr, buffer.VectorisedView{}, stack.NetworkHeaderParams{Protocol: header.ICMPv6ProtocolNumber, TTL: header.NDPHopLimit, TOS: stack.DefaultTOS}); err != nil {
sent.Dropped.Increment()
return
}
sent.NeighborAdvert.Increment()
case header.ICMPv6NeighborAdvert:
received.NeighborAdvert.Increment()
if len(v) < header.ICMPv6NeighborAdvertSize {
received.Invalid.Increment()
return
}
na := header.NDPNeighborAdvert(h.NDPPayload())
targetAddr := na.TargetAddress()
stack := r.Stack()
rxNICID := r.NICID()
isTentative, err := stack.IsAddrTentative(rxNICID, targetAddr)
if err != nil {
// We will only get an error if rxNICID is unrecognized,
// which should not happen. For now short-circuit this
// packet.
//
// TODO(b/141002840): Handle this better?
return
}
if isTentative {
// We just got an NA from a node that owns an address we
// are performing DAD on, implying the address is not
// unique. In this case we let the stack know so it can
// handle such a scenario and do nothing furthur with
// the NDP NA.
stack.DupTentativeAddrDetected(rxNICID, targetAddr)
return
}
// At this point we know that the targetAddress is not tentative
// on rxNICID. However, targetAddr may still be assigned to
// rxNICID but not tentative (it could be permanent). Such a
// scenario is beyond the scope of RFC 4862. As such, we simply
// ignore such a scenario for now and proceed as normal.
//
// TODO(b/143147598): Handle the scenario described above. Also
// inform the netstack integration that a duplicate address was
// detected outside of DAD.
e.linkAddrCache.AddLinkAddress(e.nicid, targetAddr, r.RemoteLinkAddress)
if targetAddr != r.RemoteAddress {
e.linkAddrCache.AddLinkAddress(e.nicid, r.RemoteAddress, r.RemoteLinkAddress)
}
case header.ICMPv6EchoRequest:
received.EchoRequest.Increment()
if len(v) < header.ICMPv6EchoMinimumSize {
received.Invalid.Increment()
return
}
vv.TrimFront(header.ICMPv6EchoMinimumSize)
hdr := buffer.NewPrependable(int(r.MaxHeaderLength()) + header.ICMPv6EchoMinimumSize)
pkt := header.ICMPv6(hdr.Prepend(header.ICMPv6EchoMinimumSize))
copy(pkt, h)
pkt.SetType(header.ICMPv6EchoReply)
pkt.SetChecksum(header.ICMPv6Checksum(pkt, r.LocalAddress, r.RemoteAddress, vv))
if err := r.WritePacket(nil /* gso */, hdr, vv, stack.NetworkHeaderParams{Protocol: header.ICMPv6ProtocolNumber, TTL: r.DefaultTTL(), TOS: stack.DefaultTOS}); err != nil {
sent.Dropped.Increment()
return
}
sent.EchoReply.Increment()
case header.ICMPv6EchoReply:
received.EchoReply.Increment()
if len(v) < header.ICMPv6EchoMinimumSize {
received.Invalid.Increment()
return
}
e.dispatcher.DeliverTransportPacket(r, header.ICMPv6ProtocolNumber, netHeader, vv)
case header.ICMPv6TimeExceeded:
received.TimeExceeded.Increment()
case header.ICMPv6ParamProblem:
received.ParamProblem.Increment()
case header.ICMPv6RouterSolicit:
received.RouterSolicit.Increment()
case header.ICMPv6RouterAdvert:
received.RouterAdvert.Increment()
case header.ICMPv6RedirectMsg:
received.RedirectMsg.Increment()
default:
received.Invalid.Increment()
}
}
const (
ndpSolicitedFlag = 1 << 6
ndpOverrideFlag = 1 << 5
ndpOptSrcLinkAddr = 1
ndpOptDstLinkAddr = 2
icmpV6FlagOffset = 4
icmpV6OptOffset = 24
icmpV6LengthOffset = 25
)
var broadcastMAC = tcpip.LinkAddress([]byte{0xff, 0xff, 0xff, 0xff, 0xff, 0xff})
var _ stack.LinkAddressResolver = (*protocol)(nil)
// LinkAddressProtocol implements stack.LinkAddressResolver.
func (*protocol) LinkAddressProtocol() tcpip.NetworkProtocolNumber {
return header.IPv6ProtocolNumber
}
// LinkAddressRequest implements stack.LinkAddressResolver.
func (*protocol) LinkAddressRequest(addr, localAddr tcpip.Address, linkEP stack.LinkEndpoint) *tcpip.Error {
snaddr := header.SolicitedNodeAddr(addr)
r := &stack.Route{
LocalAddress: localAddr,
RemoteAddress: snaddr,
RemoteLinkAddress: broadcastMAC,
}
hdr := buffer.NewPrependable(int(linkEP.MaxHeaderLength()) + header.IPv6MinimumSize + header.ICMPv6NeighborAdvertSize)
pkt := header.ICMPv6(hdr.Prepend(header.ICMPv6NeighborAdvertSize))
pkt.SetType(header.ICMPv6NeighborSolicit)
copy(pkt[icmpV6OptOffset-len(addr):], addr)
pkt[icmpV6OptOffset] = ndpOptSrcLinkAddr
pkt[icmpV6LengthOffset] = 1
copy(pkt[icmpV6LengthOffset+1:], linkEP.LinkAddress())
pkt.SetChecksum(header.ICMPv6Checksum(pkt, r.LocalAddress, r.RemoteAddress, buffer.VectorisedView{}))
length := uint16(hdr.UsedLength())
ip := header.IPv6(hdr.Prepend(header.IPv6MinimumSize))
ip.Encode(&header.IPv6Fields{
PayloadLength: length,
NextHeader: uint8(header.ICMPv6ProtocolNumber),
HopLimit: header.NDPHopLimit,
SrcAddr: r.LocalAddress,
DstAddr: r.RemoteAddress,
})
// TODO(stijlist): count this in ICMP stats.
return linkEP.WritePacket(r, nil /* gso */, hdr, buffer.VectorisedView{}, ProtocolNumber)
}
// ResolveStaticAddress implements stack.LinkAddressResolver.
func (*protocol) ResolveStaticAddress(addr tcpip.Address) (tcpip.LinkAddress, bool) {
if header.IsV6MulticastAddress(addr) {
// RFC 2464 Transmission of IPv6 Packets over Ethernet Networks
//
// 7. Address Mapping -- Multicast
//
// An IPv6 packet with a multicast destination address DST,
// consisting of the sixteen octets DST[1] through DST[16], is
// transmitted to the Ethernet multicast address whose first
// two octets are the value 3333 hexadecimal and whose last
// four octets are the last four octets of DST.
return tcpip.LinkAddress([]byte{
0x33,
0x33,
addr[header.IPv6AddressSize-4],
addr[header.IPv6AddressSize-3],
addr[header.IPv6AddressSize-2],
addr[header.IPv6AddressSize-1],
}), true
}
return "", false
}
|