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
|
// +build linux
package nsinit
import (
"fmt"
"os"
"runtime"
"syscall"
"github.com/dotcloud/docker/pkg/label"
"github.com/dotcloud/docker/pkg/libcontainer"
"github.com/dotcloud/docker/pkg/libcontainer/apparmor"
"github.com/dotcloud/docker/pkg/libcontainer/capabilities"
"github.com/dotcloud/docker/pkg/libcontainer/network"
"github.com/dotcloud/docker/pkg/libcontainer/utils"
"github.com/dotcloud/docker/pkg/system"
"github.com/dotcloud/docker/pkg/user"
)
// Init is the init process that first runs inside a new namespace to setup mounts, users, networking,
// and other options required for the new container.
func (ns *linuxNs) Init(container *libcontainer.Container, uncleanRootfs, console string, syncPipe *SyncPipe, args []string) error {
rootfs, err := utils.ResolveRootfs(uncleanRootfs)
if err != nil {
return err
}
// We always read this as it is a way to sync with the parent as well
ns.logger.Printf("reading from sync pipe fd %d\n", syncPipe.child.Fd())
context, err := syncPipe.ReadFromParent()
if err != nil {
syncPipe.Close()
return err
}
ns.logger.Println("received context from parent")
syncPipe.Close()
if console != "" {
ns.logger.Printf("setting up %s as console\n", console)
slave, err := system.OpenTerminal(console, syscall.O_RDWR)
if err != nil {
return fmt.Errorf("open terminal %s", err)
}
if err := dupSlave(slave); err != nil {
return fmt.Errorf("dup2 slave %s", err)
}
}
if _, err := system.Setsid(); err != nil {
return fmt.Errorf("setsid %s", err)
}
if console != "" {
if err := system.Setctty(); err != nil {
return fmt.Errorf("setctty %s", err)
}
}
if err := setupNetwork(container, context); err != nil {
return fmt.Errorf("setup networking %s", err)
}
label.Init()
ns.logger.Println("setup mount namespace")
if err := setupNewMountNamespace(rootfs, container.Mounts, console, container.ReadonlyFs, container.NoPivotRoot, container.Context["mount_label"]); err != nil {
return fmt.Errorf("setup mount namespace %s", err)
}
if err := system.Sethostname(container.Hostname); err != nil {
return fmt.Errorf("sethostname %s", err)
}
if err := finalizeNamespace(container); err != nil {
return fmt.Errorf("finalize namespace %s", err)
}
if profile := container.Context["apparmor_profile"]; profile != "" {
ns.logger.Printf("setting apparmor profile %s\n", profile)
if err := apparmor.ApplyProfile(os.Getpid(), profile); err != nil {
return err
}
}
runtime.LockOSThread()
if err := label.SetProcessLabel(container.Context["process_label"]); err != nil {
return fmt.Errorf("SetProcessLabel label %s", err)
}
ns.logger.Printf("execing %s\n", args[0])
return system.Execv(args[0], args[0:], container.Env)
}
func setupUser(container *libcontainer.Container) error {
switch container.User {
case "root", "":
if err := system.Setgroups(nil); err != nil {
return err
}
if err := system.Setresgid(0, 0, 0); err != nil {
return err
}
if err := system.Setresuid(0, 0, 0); err != nil {
return err
}
default:
uid, gid, suppGids, err := user.GetUserGroupSupplementary(container.User, syscall.Getuid(), syscall.Getgid())
if err != nil {
return err
}
if err := system.Setgroups(suppGids); err != nil {
return err
}
if err := system.Setgid(gid); err != nil {
return err
}
if err := system.Setuid(uid); err != nil {
return err
}
}
return nil
}
// dupSlave dup2 the pty slave's fd into stdout and stdin and ensures that
// the slave's fd is 0, or stdin
func dupSlave(slave *os.File) error {
if err := system.Dup2(slave.Fd(), 0); err != nil {
return err
}
if err := system.Dup2(slave.Fd(), 1); err != nil {
return err
}
if err := system.Dup2(slave.Fd(), 2); err != nil {
return err
}
return nil
}
// setupVethNetwork uses the Network config if it is not nil to initialize
// the new veth interface inside the container for use by changing the name to eth0
// setting the MTU and IP address along with the default gateway
func setupNetwork(container *libcontainer.Container, context libcontainer.Context) error {
for _, config := range container.Networks {
strategy, err := network.GetStrategy(config.Type)
if err != nil {
return err
}
err1 := strategy.Initialize(config, context)
if err1 != nil {
return err1
}
}
return nil
}
// finalizeNamespace drops the caps and sets the correct user
// and working dir before execing the command inside the namespace
func finalizeNamespace(container *libcontainer.Container) error {
if err := capabilities.DropCapabilities(container); err != nil {
return fmt.Errorf("drop capabilities %s", err)
}
if err := setupUser(container); err != nil {
return fmt.Errorf("setup user %s", err)
}
if container.WorkingDir != "" {
if err := system.Chdir(container.WorkingDir); err != nil {
return fmt.Errorf("chdir to %s %s", container.WorkingDir, err)
}
}
return nil
}
|
