summaryrefslogtreecommitdiff
path: root/libpod/networking_freebsd.go
blob: 5bc2872fc847dcf870472dd47b04897871b26336 (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
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
//go:build freebsd
// +build freebsd

package libpod

import (
	"crypto/rand"
	jdec "encoding/json"
	"errors"
	"fmt"
	"net"
	"os/exec"
	"path/filepath"
	"sort"

	"github.com/containers/buildah/pkg/jail"
	"github.com/containers/common/libnetwork/types"
	"github.com/containers/podman/v4/libpod/define"
	"github.com/containers/podman/v4/libpod/events"
	"github.com/containers/podman/v4/pkg/util"
	"github.com/containers/storage/pkg/lockfile"
	"github.com/sirupsen/logrus"
)

type Netstat struct {
	Statistics NetstatInterface `json:"statistics"`
}

type NetstatInterface struct {
	Interface []NetstatAddress `json:"interface"`
}

type NetstatAddress struct {
	Name    string `json:"name"`
	Flags   string `json:"flags"`
	Mtu     int    `json:"mtu"`
	Network string `json:"network"`
	Address string `json:"address"`

	ReceivedPackets uint64 `json:"received-packets"`
	ReceivedBytes   uint64 `json:"received-bytes"`
	ReceivedErrors  uint64 `json:"received-errors"`

	SentPackets uint64 `json:"sent-packets"`
	SentBytes   uint64 `json:"sent-bytes"`
	SentErrors  uint64 `json:"send-errors"`

	DroppedPackets uint64 `json:"dropped-packets"`

	Collisions uint64 `json:"collisions"`
}

// copied from github.com/vishvanada/netlink which does not build on freebsd
type LinkStatistics64 struct {
	RxPackets         uint64
	TxPackets         uint64
	RxBytes           uint64
	TxBytes           uint64
	RxErrors          uint64
	TxErrors          uint64
	RxDropped         uint64
	TxDropped         uint64
	Multicast         uint64
	Collisions        uint64
	RxLengthErrors    uint64
	RxOverErrors      uint64
	RxCrcErrors       uint64
	RxFrameErrors     uint64
	RxFifoErrors      uint64
	RxMissedErrors    uint64
	TxAbortedErrors   uint64
	TxCarrierErrors   uint64
	TxFifoErrors      uint64
	TxHeartbeatErrors uint64
	TxWindowErrors    uint64
	RxCompressed      uint64
	TxCompressed      uint64
}

type RootlessNetNS struct {
	dir  string
	Lock lockfile.Locker
}

// getPath will join the given path to the rootless netns dir
func (r *RootlessNetNS) getPath(path string) string {
	return filepath.Join(r.dir, path)
}

// Do - run the given function in the rootless netns.
// It does not lock the rootlessCNI lock, the caller
// should only lock when needed, e.g. for cni operations.
func (r *RootlessNetNS) Do(toRun func() error) error {
	return errors.New("not supported on freebsd")
}

// Cleanup the rootless network namespace if needed.
// It checks if we have running containers with the bridge network mode.
// Cleanup() expects that r.Lock is locked
func (r *RootlessNetNS) Cleanup(runtime *Runtime) error {
	return errors.New("not supported on freebsd")
}

// GetRootlessNetNs returns the rootless netns object. If create is set to true
// the rootless network namespace will be created if it does not exists already.
// If called as root it returns always nil.
// On success the returned RootlessCNI lock is locked and must be unlocked by the caller.
func (r *Runtime) GetRootlessNetNs(new bool) (*RootlessNetNS, error) {
	return nil, nil
}

func GetSlirp4netnsIP(subnet *net.IPNet) (*net.IP, error) {
	return nil, errors.New("not implemented GetSlirp4netnsIP")
}

// While there is code in container_internal.go which calls this, in
// my testing network creation always seems to go through createNetNS.
func (r *Runtime) setupNetNS(ctr *Container) error {
	return errors.New("not implemented (*Runtime) setupNetNS")
}

// Create and configure a new network namespace for a container
func (r *Runtime) configureNetNS(ctr *Container, ctrNS *jailNetNS) (status map[string]types.StatusBlock, rerr error) {
	if err := r.exposeMachinePorts(ctr.config.PortMappings); err != nil {
		return nil, err
	}
	defer func() {
		// make sure to unexpose the gvproxy ports when an error happens
		if rerr != nil {
			if err := r.unexposeMachinePorts(ctr.config.PortMappings); err != nil {
				logrus.Errorf("failed to free gvproxy machine ports: %v", err)
			}
		}
	}()
	networks, err := ctr.networks()
	if err != nil {
		return nil, err
	}
	// All networks have been removed from the container.
	// This is effectively forcing net=none.
	if len(networks) == 0 {
		return nil, nil
	}

	netOpts := ctr.getNetworkOptions(networks)
	netStatus, err := r.setUpNetwork(ctrNS.Name, netOpts)
	if err != nil {
		return nil, err
	}

	return netStatus, err
}

// Create and configure a new network namespace for a container
func (r *Runtime) createNetNS(ctr *Container) (n *jailNetNS, q map[string]types.StatusBlock, retErr error) {
	b := make([]byte, 16)
	_, err := rand.Reader.Read(b)
	if err != nil {
		return nil, nil, fmt.Errorf("failed to generate random vnet name: %v", err)
	}
	ctrNS := &jailNetNS{Name: fmt.Sprintf("vnet-%x-%x-%x-%x-%x", b[0:4], b[4:6], b[6:8], b[8:10], b[10:])}

	jconf := jail.NewConfig()
	jconf.Set("name", ctrNS.Name)
	jconf.Set("vnet", jail.NEW)
	jconf.Set("children.max", 1)
	jconf.Set("persist", true)
	jconf.Set("enforce_statfs", 0)
	jconf.Set("devfs_ruleset", 4)
	jconf.Set("allow.raw_sockets", true)
	jconf.Set("allow.chflags", true)
	jconf.Set("securelevel", -1)
	if _, err := jail.Create(jconf); err != nil {
		logrus.Debugf("Failed to create vnet jail %s for container %s", ctrNS.Name, ctr.ID())
	}

	logrus.Debugf("Created vnet jail %s for container %s", ctrNS.Name, ctr.ID())

	var networkStatus map[string]types.StatusBlock
	networkStatus, err = r.configureNetNS(ctr, ctrNS)
	return ctrNS, networkStatus, err
}

// Tear down a network namespace, undoing all state associated with it.
func (r *Runtime) teardownNetNS(ctr *Container) error {
	if err := r.unexposeMachinePorts(ctr.config.PortMappings); err != nil {
		// do not return an error otherwise we would prevent network cleanup
		logrus.Errorf("failed to free gvproxy machine ports: %v", err)
	}
	if err := r.teardownCNI(ctr); err != nil {
		return err
	}

	if ctr.state.NetNS != nil {
		// Rather than destroying the jail immediately, reset the
		// persist flag so that it will live until the container is
		// done.
		netjail, err := jail.FindByName(ctr.state.NetNS.Name)
		if err != nil {
			return fmt.Errorf("finding network jail %s: %w", ctr.state.NetNS.Name, err)
		}
		jconf := jail.NewConfig()
		jconf.Set("persist", false)
		if err := netjail.Set(jconf); err != nil {
			return fmt.Errorf("releasing network jail %s: %w", ctr.state.NetNS.Name, err)
		}

		ctr.state.NetNS = nil
	}

	return nil
}

func getContainerNetIO(ctr *Container) (*LinkStatistics64, error) {
	if ctr.state.NetNS == nil {
		// If NetNS is nil, it was set as none, and no netNS
		// was set up this is a valid state and thus return no
		// error, nor any statistics
		return nil, nil
	}

	// FIXME get the interface from the container netstatus
	cmd := exec.Command("jexec", ctr.state.NetNS.Name, "netstat", "-bI", "eth0", "--libxo", "json")
	out, err := cmd.Output()
	if err != nil {
		return nil, err
	}
	stats := Netstat{}
	if err := jdec.Unmarshal(out, &stats); err != nil {
		return nil, err
	}

	// Find the link stats
	for _, ifaddr := range stats.Statistics.Interface {
		if ifaddr.Mtu > 0 {
			return &LinkStatistics64{
				RxPackets:  ifaddr.ReceivedPackets,
				TxPackets:  ifaddr.SentPackets,
				RxBytes:    ifaddr.ReceivedBytes,
				TxBytes:    ifaddr.SentBytes,
				RxErrors:   ifaddr.ReceivedErrors,
				TxErrors:   ifaddr.SentErrors,
				RxDropped:  ifaddr.DroppedPackets,
				Collisions: ifaddr.Collisions,
			}, nil
		}
	}

	return &LinkStatistics64{}, nil
}

// resultToBasicNetworkConfig produces an InspectBasicNetworkConfig from a CNI
// result
func resultToBasicNetworkConfig(result types.StatusBlock) define.InspectBasicNetworkConfig {
	config := define.InspectBasicNetworkConfig{}
	interfaceNames := make([]string, 0, len(result.Interfaces))
	for interfaceName := range result.Interfaces {
		interfaceNames = append(interfaceNames, interfaceName)
	}
	// ensure consistent inspect results by sorting
	sort.Strings(interfaceNames)
	for _, interfaceName := range interfaceNames {
		netInt := result.Interfaces[interfaceName]
		for _, netAddress := range netInt.Subnets {
			size, _ := netAddress.IPNet.Mask.Size()
			if netAddress.IPNet.IP.To4() != nil {
				// ipv4
				if config.IPAddress == "" {
					config.IPAddress = netAddress.IPNet.IP.String()
					config.IPPrefixLen = size
					config.Gateway = netAddress.Gateway.String()
				} else {
					config.SecondaryIPAddresses = append(config.SecondaryIPAddresses, define.Address{Addr: netAddress.IPNet.IP.String(), PrefixLength: size})
				}
			} else {
				// ipv6
				if config.GlobalIPv6Address == "" {
					config.GlobalIPv6Address = netAddress.IPNet.IP.String()
					config.GlobalIPv6PrefixLen = size
					config.IPv6Gateway = netAddress.Gateway.String()
				} else {
					config.SecondaryIPv6Addresses = append(config.SecondaryIPv6Addresses, define.Address{Addr: netAddress.IPNet.IP.String(), PrefixLength: size})
				}
			}
		}
		if config.MacAddress == "" {
			config.MacAddress = netInt.MacAddress.String()
		} else {
			config.AdditionalMacAddresses = append(config.AdditionalMacAddresses, netInt.MacAddress.String())
		}
	}
	return config
}

// NetworkDisconnect removes a container from the network
func (c *Container) NetworkDisconnect(nameOrID, netName string, force bool) error {
	// only the bridge mode supports cni networks
	if err := isBridgeNetMode(c.config.NetMode); err != nil {
		return err
	}

	c.lock.Lock()
	defer c.lock.Unlock()

	networks, err := c.networks()
	if err != nil {
		return err
	}

	// check if network exists and if the input is a ID we get the name
	// CNI only uses names so it is important that we only use the name
	netName, err = c.runtime.normalizeNetworkName(netName)
	if err != nil {
		return err
	}

	_, nameExists := networks[netName]
	if !nameExists && len(networks) > 0 {
		return fmt.Errorf("container %s is not connected to network %s", nameOrID, netName)
	}

	if err := c.syncContainer(); err != nil {
		return err
	}
	// get network status before we disconnect
	networkStatus := c.getNetworkStatus()

	if err := c.runtime.state.NetworkDisconnect(c, netName); err != nil {
		return err
	}

	c.newNetworkEvent(events.NetworkDisconnect, netName)
	if !c.ensureState(define.ContainerStateRunning, define.ContainerStateCreated) {
		return nil
	}

	opts := types.NetworkOptions{
		ContainerID:   c.config.ID,
		ContainerName: getCNIPodName(c),
	}
	opts.PortMappings = c.convertPortMappings()
	opts.Networks = map[string]types.PerNetworkOptions{
		netName: networks[netName],
	}

	// update network status if container is running
	oldStatus, statusExist := networkStatus[netName]
	delete(networkStatus, netName)
	c.state.NetworkStatus = networkStatus
	err = c.save()
	if err != nil {
		return err
	}

	// Update resolv.conf if required
	if statusExist {
		stringIPs := make([]string, 0, len(oldStatus.DNSServerIPs))
		for _, ip := range oldStatus.DNSServerIPs {
			stringIPs = append(stringIPs, ip.String())
		}
		if len(stringIPs) == 0 {
			return nil
		}
		logrus.Debugf("Removing DNS Servers %v from resolv.conf", stringIPs)
		if err := c.removeNameserver(stringIPs); err != nil {
			return err
		}
	}

	return nil
}

// ConnectNetwork connects a container to a given network
func (c *Container) NetworkConnect(nameOrID, netName string, netOpts types.PerNetworkOptions) error {
	// only the bridge mode supports cni networks
	if err := isBridgeNetMode(c.config.NetMode); err != nil {
		return err
	}

	c.lock.Lock()
	defer c.lock.Unlock()

	networks, err := c.networks()
	if err != nil {
		return err
	}

	// check if network exists and if the input is a ID we get the name
	// CNI only uses names so it is important that we only use the name
	netName, err = c.runtime.normalizeNetworkName(netName)
	if err != nil {
		return err
	}

	if err := c.syncContainer(); err != nil {
		return err
	}

	// get network status before we connect
	networkStatus := c.getNetworkStatus()

	// always add the short id as alias for docker compat
	netOpts.Aliases = append(netOpts.Aliases, c.config.ID[:12])

	if netOpts.InterfaceName == "" {
		netOpts.InterfaceName = getFreeInterfaceName(networks)
		if netOpts.InterfaceName == "" {
			return errors.New("could not find free network interface name")
		}
	}

	if err := c.runtime.state.NetworkConnect(c, netName, netOpts); err != nil {
		return err
	}
	c.newNetworkEvent(events.NetworkConnect, netName)
	if !c.ensureState(define.ContainerStateRunning, define.ContainerStateCreated) {
		return nil
	}

	opts := types.NetworkOptions{
		ContainerID:   c.config.ID,
		ContainerName: getCNIPodName(c),
	}
	opts.PortMappings = c.convertPortMappings()
	opts.Networks = map[string]types.PerNetworkOptions{
		netName: netOpts,
	}

	/*
		results, err := c.runtime.setUpNetwork(c.state.NetNS.Path(), opts)
		if err != nil {
			return err
		}
		if len(results) != 1 {
			return errors.New("when adding aliases, results must be of length 1")
		}
	*/
	var results map[string]types.StatusBlock

	// update network status
	if networkStatus == nil {
		networkStatus = make(map[string]types.StatusBlock, 1)
	}
	networkStatus[netName] = results[netName]
	c.state.NetworkStatus = networkStatus

	err = c.save()
	if err != nil {
		return err
	}

	// The first network needs a port reload to set the correct child ip for the rootlessport process.
	// Adding a second network does not require a port reload because the child ip is still valid.

	ipv6, err := c.checkForIPv6(networkStatus)
	if err != nil {
		return err
	}

	// Update resolv.conf if required
	stringIPs := make([]string, 0, len(results[netName].DNSServerIPs))
	for _, ip := range results[netName].DNSServerIPs {
		if (ip.To4() == nil) && !ipv6 {
			continue
		}
		stringIPs = append(stringIPs, ip.String())
	}
	if len(stringIPs) == 0 {
		return nil
	}
	logrus.Debugf("Adding DNS Servers %v to resolv.conf", stringIPs)
	if err := c.addNameserver(stringIPs); err != nil {
		return err
	}

	return nil
}

// get a free interface name for a new network
// return an empty string if no free name was found
func getFreeInterfaceName(networks map[string]types.PerNetworkOptions) string {
	ifNames := make([]string, 0, len(networks))
	for _, opts := range networks {
		ifNames = append(ifNames, opts.InterfaceName)
	}
	for i := 0; i < 100000; i++ {
		ifName := fmt.Sprintf("eth%d", i)
		if !util.StringInSlice(ifName, ifNames) {
			return ifName
		}
	}
	return ""
}

// DisconnectContainerFromNetwork removes a container from its CNI network
func (r *Runtime) DisconnectContainerFromNetwork(nameOrID, netName string, force bool) error {
	ctr, err := r.LookupContainer(nameOrID)
	if err != nil {
		return err
	}
	return ctr.NetworkDisconnect(nameOrID, netName, force)
}

// ConnectContainerToNetwork connects a container to a CNI network
func (r *Runtime) ConnectContainerToNetwork(nameOrID, netName string, netOpts types.PerNetworkOptions) error {
	ctr, err := r.LookupContainer(nameOrID)
	if err != nil {
		return err
	}
	return ctr.NetworkConnect(nameOrID, netName, netOpts)
}

// normalizeNetworkName takes a network name, a partial or a full network ID and returns the network name.
// If the network is not found a errors is returned.
func (r *Runtime) normalizeNetworkName(nameOrID string) (string, error) {
	net, err := r.network.NetworkInspect(nameOrID)
	if err != nil {
		return "", err
	}
	return net.Name, nil
}

// ocicniPortsToNetTypesPorts convert the old port format to the new one
// while deduplicating ports into ranges
func ocicniPortsToNetTypesPorts(ports []types.OCICNIPortMapping) []types.PortMapping {
	if len(ports) == 0 {
		return nil
	}

	newPorts := make([]types.PortMapping, 0, len(ports))

	// first sort the ports
	sort.Slice(ports, func(i, j int) bool {
		return compareOCICNIPorts(ports[i], ports[j])
	})

	// we already check if the slice is empty so we can use the first element
	currentPort := types.PortMapping{
		HostIP:        ports[0].HostIP,
		HostPort:      uint16(ports[0].HostPort),
		ContainerPort: uint16(ports[0].ContainerPort),
		Protocol:      ports[0].Protocol,
		Range:         1,
	}

	for i := 1; i < len(ports); i++ {
		if ports[i].HostIP == currentPort.HostIP &&
			ports[i].Protocol == currentPort.Protocol &&
			ports[i].HostPort-int32(currentPort.Range) == int32(currentPort.HostPort) &&
			ports[i].ContainerPort-int32(currentPort.Range) == int32(currentPort.ContainerPort) {
			currentPort.Range = currentPort.Range + 1
		} else {
			newPorts = append(newPorts, currentPort)
			currentPort = types.PortMapping{
				HostIP:        ports[i].HostIP,
				HostPort:      uint16(ports[i].HostPort),
				ContainerPort: uint16(ports[i].ContainerPort),
				Protocol:      ports[i].Protocol,
				Range:         1,
			}
		}
	}
	newPorts = append(newPorts, currentPort)
	return newPorts
}

// compareOCICNIPorts will sort the ocicni ports by
// 1) host ip
// 2) protocol
// 3) hostPort
// 4) container port
func compareOCICNIPorts(i, j types.OCICNIPortMapping) bool {
	if i.HostIP != j.HostIP {
		return i.HostIP < j.HostIP
	}

	if i.Protocol != j.Protocol {
		return i.Protocol < j.Protocol
	}

	if i.HostPort != j.HostPort {
		return i.HostPort < j.HostPort
	}

	return i.ContainerPort < j.ContainerPort
}

func (c *Container) setupRootlessNetwork() error {
	return nil
}