summaryrefslogtreecommitdiffhomepage
path: root/pkg/sentry/socket/epsocket/provider.go
blob: 0a9dfa6c3289e2d05fd5e66d0bc9f8bad44a336d (plain)
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
// 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 epsocket

import (
	"syscall"

	"gvisor.dev/gvisor/pkg/abi/linux"
	"gvisor.dev/gvisor/pkg/sentry/context"
	"gvisor.dev/gvisor/pkg/sentry/fs"
	"gvisor.dev/gvisor/pkg/sentry/kernel"
	"gvisor.dev/gvisor/pkg/sentry/kernel/auth"
	"gvisor.dev/gvisor/pkg/sentry/socket"
	"gvisor.dev/gvisor/pkg/syserr"
	"gvisor.dev/gvisor/pkg/tcpip"
	"gvisor.dev/gvisor/pkg/tcpip/header"
	"gvisor.dev/gvisor/pkg/tcpip/network/ipv4"
	"gvisor.dev/gvisor/pkg/tcpip/network/ipv6"
	"gvisor.dev/gvisor/pkg/tcpip/transport/tcp"
	"gvisor.dev/gvisor/pkg/tcpip/transport/udp"
	"gvisor.dev/gvisor/pkg/waiter"
)

// provider is an inet socket provider.
type provider struct {
	family   int
	netProto tcpip.NetworkProtocolNumber
}

// getTransportProtocol figures out transport protocol. Currently only TCP,
// UDP, and ICMP are supported. The bool return value is true when this socket
// is associated with a transport protocol. This is only false for SOCK_RAW,
// IPPROTO_IP sockets.
func getTransportProtocol(ctx context.Context, stype linux.SockType, protocol int) (tcpip.TransportProtocolNumber, bool, *syserr.Error) {
	switch stype {
	case linux.SOCK_STREAM:
		if protocol != 0 && protocol != syscall.IPPROTO_TCP {
			return 0, true, syserr.ErrInvalidArgument
		}
		return tcp.ProtocolNumber, true, nil

	case linux.SOCK_DGRAM:
		switch protocol {
		case 0, syscall.IPPROTO_UDP:
			return udp.ProtocolNumber, true, nil
		case syscall.IPPROTO_ICMP:
			return header.ICMPv4ProtocolNumber, true, nil
		case syscall.IPPROTO_ICMPV6:
			return header.ICMPv6ProtocolNumber, true, nil
		}

	case linux.SOCK_RAW:
		// Raw sockets require CAP_NET_RAW.
		creds := auth.CredentialsFromContext(ctx)
		if !creds.HasCapability(linux.CAP_NET_RAW) {
			return 0, true, syserr.ErrNotPermitted
		}

		switch protocol {
		case syscall.IPPROTO_ICMP:
			return header.ICMPv4ProtocolNumber, true, nil
		case syscall.IPPROTO_UDP:
			return header.UDPProtocolNumber, true, nil
		case syscall.IPPROTO_TCP:
			return header.TCPProtocolNumber, true, nil
		// IPPROTO_RAW signifies that the raw socket isn't assigned to
		// a transport protocol. Users will be able to write packets'
		// IP headers and won't receive anything.
		case syscall.IPPROTO_RAW:
			return tcpip.TransportProtocolNumber(0), false, nil
		}
	}
	return 0, true, syserr.ErrProtocolNotSupported
}

// Socket creates a new socket object for the AF_INET or AF_INET6 family.
func (p *provider) Socket(t *kernel.Task, stype linux.SockType, protocol int) (*fs.File, *syserr.Error) {
	// Fail right away if we don't have a stack.
	stack := t.NetworkContext()
	if stack == nil {
		// Don't propagate an error here. Instead, allow the socket
		// code to continue searching for another provider.
		return nil, nil
	}
	eps, ok := stack.(*Stack)
	if !ok {
		return nil, nil
	}

	// Figure out the transport protocol.
	transProto, associated, err := getTransportProtocol(t, stype, protocol)
	if err != nil {
		return nil, err
	}

	// Create the endpoint.
	var ep tcpip.Endpoint
	var e *tcpip.Error
	wq := &waiter.Queue{}
	if stype == linux.SOCK_RAW {
		ep, e = eps.Stack.NewRawEndpoint(transProto, p.netProto, wq, associated)
	} else {
		ep, e = eps.Stack.NewEndpoint(transProto, p.netProto, wq)
	}
	if e != nil {
		return nil, syserr.TranslateNetstackError(e)
	}

	return New(t, p.family, stype, int(transProto), wq, ep)
}

// Pair just returns nil sockets (not supported).
func (*provider) Pair(*kernel.Task, linux.SockType, int) (*fs.File, *fs.File, *syserr.Error) {
	return nil, nil, nil
}

// init registers socket providers for AF_INET and AF_INET6.
func init() {
	// Providers backed by netstack.
	p := []provider{
		{
			family:   linux.AF_INET,
			netProto: ipv4.ProtocolNumber,
		},
		{
			family:   linux.AF_INET6,
			netProto: ipv6.ProtocolNumber,
		},
	}

	for i := range p {
		socket.RegisterProvider(p[i].family, &p[i])
	}
}