summaryrefslogtreecommitdiffstats
path: root/lib/netns_linux.c
blob: 297e2c5952dfd1124720778d52d1d9b4423597bc (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
591
592
593
594
595
596
597
// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * NS functions.
 * Copyright (C) 2014 6WIND S.A.
 */

#include <zebra.h>

#ifdef HAVE_NETNS
#undef _GNU_SOURCE
#define _GNU_SOURCE

#include <sched.h>
#endif

/* for basename */
#include <libgen.h>

#include "if.h"
#include "ns.h"
#include "log.h"
#include "memory.h"
#include "command.h"
#include "vty.h"
#include "vrf.h"
#include "lib_errors.h"

DEFINE_MTYPE_STATIC(LIB, NS, "NetNS Context");
DEFINE_MTYPE_STATIC(LIB, NS_NAME, "NetNS Name");

static inline int ns_compare(const struct ns *ns, const struct ns *ns2);
static struct ns *ns_lookup_name_internal(const char *name);

RB_GENERATE(ns_head, ns, entry, ns_compare)

static struct ns_head ns_tree = RB_INITIALIZER(&ns_tree);

static struct ns *default_ns;
static int ns_current_ns_fd;
static int ns_default_ns_fd;

static int ns_debug;

struct ns_map_nsid {
	RB_ENTRY(ns_map_nsid) id_entry;
	ns_id_t ns_id_external;
	ns_id_t ns_id;
};

static inline int ns_map_compare(const struct ns_map_nsid *a,
				   const struct ns_map_nsid *b)
{
	return (a->ns_id - b->ns_id);
}

RB_HEAD(ns_map_nsid_head, ns_map_nsid);
RB_PROTOTYPE(ns_map_nsid_head, ns_map_nsid, id_entry, ns_map_compare);
RB_GENERATE(ns_map_nsid_head, ns_map_nsid, id_entry, ns_map_compare);
static struct ns_map_nsid_head ns_map_nsid_list =
		RB_INITIALIZER(&ns_map_nsid_list);

static ns_id_t ns_id_external_numbering;


#ifndef CLONE_NEWNET
#define CLONE_NEWNET 0x40000000
/* New network namespace (lo, device, names sockets, etc) */
#endif

#ifndef HAVE_SETNS
static inline int setns(int fd, int nstype)
{
#ifdef __NR_setns
	return syscall(__NR_setns, fd, nstype);
#else
	errno = EINVAL;
	return -1;
#endif
}
#endif /* !HAVE_SETNS */

#ifdef HAVE_NETNS
static int have_netns_enabled = -1;
#endif /* HAVE_NETNS */

static int have_netns(void)
{
#ifdef HAVE_NETNS
	if (have_netns_enabled < 0) {
		int fd = open(NS_DEFAULT_NAME, O_RDONLY);

		if (fd < 0)
			have_netns_enabled = 0;
		else {
			have_netns_enabled = 1;
			close(fd);
		}
	}
	return have_netns_enabled;
#else
	return 0;
#endif
}

/* Holding NS hooks  */
static struct ns_master {
	int (*ns_new_hook)(struct ns *ns);
	int (*ns_delete_hook)(struct ns *ns);
	int (*ns_enable_hook)(struct ns *ns);
	int (*ns_disable_hook)(struct ns *ns);
} ns_master = {
	0,
};

static int ns_is_enabled(struct ns *ns);

static inline int ns_compare(const struct ns *a, const struct ns *b)
{
	return (a->ns_id - b->ns_id);
}

/* Look up a NS by identifier. */
static struct ns *ns_lookup_internal(ns_id_t ns_id)
{
	struct ns ns;

	ns.ns_id = ns_id;
	return RB_FIND(ns_head, &ns_tree, &ns);
}

/* Look up a NS by name */
static struct ns *ns_lookup_name_internal(const char *name)
{
	struct ns *ns = NULL;

	RB_FOREACH (ns, ns_head, &ns_tree) {
		if (ns->name != NULL) {
			if (strcmp(name, ns->name) == 0)
				return ns;
		}
	}
	return NULL;
}

static struct ns *ns_get_created_internal(struct ns *ns, char *name,
					  ns_id_t ns_id)
{
	int created = 0;
	/*
	 * Initialize interfaces.
	 */
	if (!ns && !name && ns_id != NS_UNKNOWN)
		ns = ns_lookup_internal(ns_id);
	if (!ns && name)
		ns = ns_lookup_name_internal(name);
	if (!ns) {
		ns = XCALLOC(MTYPE_NS, sizeof(struct ns));
		ns->ns_id = ns_id;
		if (name)
			ns->name = XSTRDUP(MTYPE_NS_NAME, name);
		ns->fd = -1;
		RB_INSERT(ns_head, &ns_tree, ns);
		created = 1;
	}
	if (ns_id != ns->ns_id) {
		RB_REMOVE(ns_head, &ns_tree, ns);
		ns->ns_id = ns_id;
		RB_INSERT(ns_head, &ns_tree, ns);
	}
	if (!created)
		return ns;
	if (ns_debug) {
		if (ns->ns_id != NS_UNKNOWN)
			zlog_info("NS %u is created.", ns->ns_id);
		else
			zlog_info("NS %s is created.", ns->name);
	}
	if (ns_master.ns_new_hook)
		(*ns_master.ns_new_hook)(ns);
	return ns;
}

/*
 * Enable a NS - that is, let the NS be ready to use.
 * The NS_ENABLE_HOOK callback will be called to inform
 * that they can allocate resources in this NS.
 *
 * RETURN: 1 - enabled successfully; otherwise, 0.
 */
static int ns_enable_internal(struct ns *ns, void (*func)(ns_id_t, void *))
{
	if (!ns_is_enabled(ns)) {
		if (have_netns()) {
			ns->fd = open(ns->name, O_RDONLY);
		} else {
			ns->fd = -2;
			/* Remember ns_enable_hook has been called */
			errno = -ENOTSUP;
		}

		if (!ns_is_enabled(ns)) {
			flog_err_sys(EC_LIB_SYSTEM_CALL,
				     "Can not enable NS %u: %s!", ns->ns_id,
				     safe_strerror(errno));
			return 0;
		}

		/* Non default NS. leave */
		if (ns->ns_id == NS_UNKNOWN) {
			flog_err(EC_LIB_NS,
				 "Can not enable NS %s %u: Invalid NSID",
				 ns->name, ns->ns_id);
			return 0;
		}
		if (func)
			func(ns->ns_id, (void *)ns->vrf_ctxt);
		if (ns_debug) {
			if (have_netns())
				zlog_info("NS %u is associated with NETNS %s.",
					  ns->ns_id, ns->name);
			zlog_info("NS %u is enabled.", ns->ns_id);
		}
		/* zebra first receives NS enable event,
		 * then VRF enable event
		 */
		if (ns_master.ns_enable_hook)
			(*ns_master.ns_enable_hook)(ns);
	}

	return 1;
}

/*
 * Check whether the NS is enabled - that is, whether the NS
 * is ready to allocate resources. Currently there's only one
 * type of resource: socket.
 */
static int ns_is_enabled(struct ns *ns)
{
	if (have_netns())
		return ns && ns->fd >= 0;
	else
		return ns && ns->fd == -2 && ns->ns_id == NS_DEFAULT;
}

/*
 * Disable a NS - that is, let the NS be unusable.
 * The NS_DELETE_HOOK callback will be called to inform
 * that they must release the resources in the NS.
 */
static void ns_disable_internal(struct ns *ns)
{
	if (ns_is_enabled(ns)) {
		if (ns_debug)
			zlog_info("NS %u is to be disabled.", ns->ns_id);

		if (ns_master.ns_disable_hook)
			(*ns_master.ns_disable_hook)(ns);

		if (have_netns())
			close(ns->fd);

		ns->fd = -1;
	}
}

/* VRF list existence check by name. */
static struct ns_map_nsid *ns_map_nsid_lookup_by_nsid(ns_id_t ns_id)
{
	struct ns_map_nsid ns_map;

	ns_map.ns_id = ns_id;
	return RB_FIND(ns_map_nsid_head, &ns_map_nsid_list, &ns_map);
}

ns_id_t ns_map_nsid_with_external(ns_id_t ns_id, bool map)
{
	struct ns_map_nsid *ns_map;
	vrf_id_t ns_id_external;

	ns_map = ns_map_nsid_lookup_by_nsid(ns_id);
	if (ns_map && !map) {
		ns_id_external = ns_map->ns_id_external;
		RB_REMOVE(ns_map_nsid_head, &ns_map_nsid_list, ns_map);
		return ns_id_external;
	}
	if (ns_map)
		return ns_map->ns_id_external;
	ns_map = XCALLOC(MTYPE_NS, sizeof(struct ns_map_nsid));
	/* increase vrf_id
	 * default vrf is the first one : 0
	 */
	ns_map->ns_id_external = ns_id_external_numbering++;
	ns_map->ns_id = ns_id;
	RB_INSERT(ns_map_nsid_head, &ns_map_nsid_list, ns_map);
	return ns_map->ns_id_external;
}

struct ns *ns_get_created(struct ns *ns, char *name, ns_id_t ns_id)
{
	return ns_get_created_internal(ns, name, ns_id);
}

int ns_have_netns(void)
{
	return have_netns();
}

/* Delete a NS. This is called in ns_terminate(). */
void ns_delete(struct ns *ns)
{
	if (ns_debug)
		zlog_info("NS %u is to be deleted.", ns->ns_id);

	ns_disable(ns);

	if (ns_master.ns_delete_hook)
		(*ns_master.ns_delete_hook)(ns);

	/*
	 * I'm not entirely sure if the vrf->iflist
	 * needs to be moved into here or not.
	 */
	// if_terminate (&ns->iflist);

	RB_REMOVE(ns_head, &ns_tree, ns);
	XFREE(MTYPE_NS_NAME, ns->name);

	XFREE(MTYPE_NS, ns);
}

/* Look up the data pointer of the specified VRF. */
void *ns_info_lookup(ns_id_t ns_id)
{
	struct ns *ns = ns_lookup_internal(ns_id);

	return ns ? ns->info : NULL;
}

/* Look up a NS by name */
struct ns *ns_lookup_name(const char *name)
{
	return ns_lookup_name_internal(name);
}

int ns_enable(struct ns *ns, void (*func)(ns_id_t, void *))
{
	return ns_enable_internal(ns, func);
}

void ns_disable(struct ns *ns)
{
	ns_disable_internal(ns);
}

struct ns *ns_lookup(ns_id_t ns_id)
{
	return ns_lookup_internal(ns_id);
}

void ns_walk_func(int (*func)(struct ns *,
			      void *param_in,
			      void **param_out),
		  void *param_in,
		  void **param_out)
{
	struct ns *ns = NULL;
	int ret;

	RB_FOREACH (ns, ns_head, &ns_tree) {
		ret = func(ns, param_in, param_out);
		if (ret == NS_WALK_STOP)
			return;
	}
}

const char *ns_get_name(struct ns *ns)
{
	if (!ns)
		return NULL;
	return ns->name;
}

/* Add a NS hook. Please add hooks before calling ns_init(). */
void ns_add_hook(int type, int (*func)(struct ns *))
{
	switch (type) {
	case NS_NEW_HOOK:
		ns_master.ns_new_hook = func;
		break;
	case NS_DELETE_HOOK:
		ns_master.ns_delete_hook = func;
		break;
	case NS_ENABLE_HOOK:
		ns_master.ns_enable_hook = func;
		break;
	case NS_DISABLE_HOOK:
		ns_master.ns_disable_hook = func;
		break;
	default:
		break;
	}
}

/*
 * NS realization with NETNS
 */

char *ns_netns_pathname(struct vty *vty, const char *name)
{
	static char pathname[PATH_MAX];
	char *result;
	char *check_base;

	if (name[0] == '/') /* absolute pathname */
		result = realpath(name, pathname);
	else {
		/* relevant pathname */
		char tmp_name[PATH_MAX];

		snprintf(tmp_name, sizeof(tmp_name), "%s/%s", NS_RUN_DIR, name);
		result = realpath(tmp_name, pathname);
	}

	if (!result) {
		if (vty)
			vty_out(vty, "Invalid pathname for %s: %s\n",
				pathname,
				safe_strerror(errno));
		else
			flog_warn(EC_LIB_LINUX_NS,
				  "Invalid pathname for %s: %s", pathname,
				  safe_strerror(errno));
		return NULL;
	}
	check_base = basename(pathname);
	if (check_base != NULL && strlen(check_base) + 1 > NS_NAMSIZ) {
		if (vty)
			vty_out(vty, "NS name (%s) invalid: too long (>%d)\n",
				check_base, NS_NAMSIZ - 1);
		else
			flog_warn(EC_LIB_LINUX_NS,
				  "NS name (%s) invalid: too long (>%d)",
				  check_base, NS_NAMSIZ - 1);
		return NULL;
	}
	return pathname;
}

void ns_init(void)
{
	static int ns_initialised;

	ns_debug = 0;
	/* silently return as initialisation done */
	if (ns_initialised == 1)
		return;
	errno = 0;
	if (have_netns())
		ns_default_ns_fd = open(NS_DEFAULT_NAME, O_RDONLY);
	else {
		ns_default_ns_fd = -1;
		default_ns = NULL;
	}
	ns_current_ns_fd = -1;
	ns_initialised = 1;
}

/* Initialize NS module. */
void ns_init_management(ns_id_t default_ns_id, ns_id_t internal_ns)
{
	int fd;

	ns_init();
	default_ns = ns_get_created_internal(NULL, NULL, default_ns_id);
	if (!default_ns) {
		flog_err(EC_LIB_NS, "%s: failed to create the default NS!",
			 __func__);
		exit(1);
	}
	if (have_netns()) {
		fd = open(NS_DEFAULT_NAME, O_RDONLY);
		default_ns->fd = fd;
	}
	default_ns->internal_ns_id = internal_ns;

	/* Set the default NS name. */
	default_ns->name = XSTRDUP(MTYPE_NS_NAME, NS_DEFAULT_NAME);
	if (ns_debug)
		zlog_info("%s: default NSID is %u", __func__,
			  default_ns->ns_id);

	/* Enable the default NS. */
	if (!ns_enable(default_ns, NULL)) {
		flog_err(EC_LIB_NS, "%s: failed to enable the default NS!",
			 __func__);
		exit(1);
	}
}

/* Terminate NS module. */
void ns_terminate(void)
{
	struct ns *ns;

	while (!RB_EMPTY(ns_head, &ns_tree)) {
		ns = RB_ROOT(ns_head, &ns_tree);

		ns_delete(ns);
	}
}

int ns_switch_to_netns(const char *name)
{
	int ret;
	int fd;

	if (name == NULL)
		return -1;
	if (ns_default_ns_fd == -1)
		return -1;
	fd = open(name, O_RDONLY);
	if (fd == -1) {
		errno = EINVAL;
		return -1;
	}
	ret = setns(fd, CLONE_NEWNET);
	ns_current_ns_fd = fd;
	close(fd);
	return ret;
}

/* returns 1 if switch() was not called before
 * return status of setns() otherwise
 */
int ns_switchback_to_initial(void)
{
	if (ns_current_ns_fd != -1 && ns_default_ns_fd != -1) {
		int ret;

		ret = setns(ns_default_ns_fd, CLONE_NEWNET);
		ns_current_ns_fd = -1;
		return ret;
	}
	/* silently ignore if setns() is not called */
	return 1;
}

/* Create a socket for the NS. */
int ns_socket(int domain, int type, int protocol, ns_id_t ns_id)
{
	struct ns *ns = ns_lookup(ns_id);
	int ret;

	if (!ns || !ns_is_enabled(ns)) {
		errno = EINVAL;
		return -1;
	}
	if (have_netns()) {
		ret = (ns_id != NS_DEFAULT) ? setns(ns->fd, CLONE_NEWNET) : 0;
		if (ret >= 0) {
			ret = socket(domain, type, protocol);
			if (ns_id != NS_DEFAULT) {
				setns(ns_lookup(NS_DEFAULT)->fd, CLONE_NEWNET);
				ns_current_ns_fd = ns_id;
			}
		}
	} else
		ret = socket(domain, type, protocol);

	return ret;
}

/* if relative link_nsid matches default netns,
 * then return default absolute netns value
 * otherwise, return NS_UNKNOWN
 */
ns_id_t ns_id_get_absolute(ns_id_t ns_id_reference, ns_id_t link_nsid)
{
	struct ns *ns;

	ns = ns_lookup(ns_id_reference);
	if (!ns)
		return NS_UNKNOWN;

	if (ns->relative_default_ns != link_nsid)
		return NS_UNKNOWN;

	ns = ns_get_default();
	assert(ns);
	return ns->ns_id;
}

struct ns *ns_get_default(void)
{
	return default_ns;
}