File: virtio_vhostuser.c

package info (click to toggle)
virtio-forwarder 1.1.99.51-1~unstable
  • links: PTS
  • area: main
  • in suites: buster
  • size: 600 kB
  • sloc: ansic: 6,594; python: 921; makefile: 330; sh: 80
file content (843 lines) | stat: -rw-r--r-- 23,657 bytes parent folder | download
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
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
/*
 *   BSD LICENSE
 *
 *   Copyright(c) 2016-2017 Netronome.
 *   All rights reserved.
 *
 *   Redistribution and use in source and binary forms, with or without
 *   modification, are permitted provided that the following conditions
 *   are met:
 *
 *     * Redistributions of source code must retain the above copyright
 *       notice, this list of conditions and the following disclaimer.
 *     * Redistributions in binary form must reproduce the above copyright
 *       notice, this list of conditions and the following disclaimer in
 *       the documentation and/or other materials provided with the
 *       distribution.
 *     * Neither the name of Netronome nor the names of its
 *       contributors may be used to endorse or promote products derived
 *       from this software without specific prior written permission.
 *
 *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <pthread.h>
#include <string.h>
#include <unistd.h>
#include <sys/time.h>
#include <sys/stat.h>
#include <stdbool.h>
#include "virtio_vhostuser.h"
#define __MODULE__ "virtio_vhostuser"
#include "log.h"
#include "virtio_worker.h"
#include "ugid.h"
#include <rte_version.h>
#if RTE_VERSION >= RTE_VERSION_NUM(17,5,0,0)
#include <rte_vhost.h>
#include <linux/virtio_net.h>
#else
#include <rte_virtio_net.h>
#endif

const char* const virtio_features[] = {
	"CSUM", // 0
	"GUEST_CSUM", //1
	"CTRL_GUEST_OFFLOADS", //2
	"3",//3
	"4",//4
	"MAC",//5
	"6",//6
	"GUEST_TSO4",//7
	"GUEST_TSO6",//8
	"GUEST_ECN",//9
	"GUEST_UFO",//10
	"HOST_TSO4",//11
	"HOST_TSO6",//12
	"HOST_ECN",//13
	"HOST_UFO",//14
	"MRG_RXBUF",//15
	"STATUS",//16
	"CTRL_VQ",//17
	"CTRL_RX",//18
	"CTRL_VLAN",//19
	"20",//20
	"GUEST_ANNOUNCE",//21
	"MQ",//22
	"CTRL_MAC_ADDR",//23
	"NOTIFY_ON_EMPTY",//24
	"25",//25
	"VHOST_LOG_ALL",//26
	"ANY_LAYOUT",//27
	"RING_INDIRECT_DESC",//28
	"29",//29
	"VHOST_USER_PROTOCOL_FEATURES",//30
	"31",//31
	"VIRTIO_1.0",//VERSION_1
};

const char* const vhost_features[] = {
	"MQ",//0
	"LOG_SHMFD",//1
	"RARP"
};


#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
static pthread_t vhostuser_thread;
#endif
static char vhost_socket_name_prefix[128];
static char vhost_socket_name_suffix[128];
static bool mk_default_sockets;
struct virtio_vhostuser_conf g_vio_worker_conf;
static char *relay_ifname_map[MAX_RELAYS];

static const char
*virtio_vhostuser_id_to_name(unsigned id, char *name, size_t namesz)
{
	snprintf(name, namesz, "%s%u%s", vhost_socket_name_prefix, id,
			vhost_socket_name_suffix);

	return name;
}

/*
 * Expect name to const of vhost_socket_name_prefix followed by unsigned int id
 * followed by vhost_socket_name_suffix.
 * Return id or -1 on mismatch
 */
__attribute__ ((unused))
static int virtio_vhostuser_name_to_id(char *name)
{
	unsigned namelen = strlen(name);
	unsigned prefixlen = strlen(vhost_socket_name_prefix);
	unsigned suffixlen = strlen(vhost_socket_name_suffix);
	int spn;
	unsigned id;

	if (namelen < (prefixlen + suffixlen + 1))
		return -1;
	if (strncmp(name, vhost_socket_name_prefix, prefixlen) != 0)
		return -1;
	if (suffixlen && strcmp(name + namelen - suffixlen,
			vhost_socket_name_suffix) != 0)
		return -1;

	spn = strspn(name + prefixlen, "0123456789");
	if (namelen < prefixlen + spn)
		return -1;
	id = atoi(name + prefixlen);
	if (id < MAX_RELAYS)
		return id;

	return -1;
}

static int get_relay_for_sock(const char *vhost_path)
{
	for (int i=0; i<MAX_RELAYS; ++i) {
		if (relay_ifname_map[i] &&
				strcmp(relay_ifname_map[i], vhost_path) == 0)
			return i;
	}

	return -1;
}

#if RTE_VERSION >= RTE_VERSION_NUM(16,7,0,0)
static int virtio_vhostuser_new_device_cb(int vid)
{
	int id;
	char ifname[128];
#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
	unsigned num_vring_pairs = rte_vhost_get_queue_num(vid);
#else
	unsigned num_vring_pairs = rte_vhost_get_vring_num(vid)/2;
#endif

	rte_vhost_get_ifname(vid, ifname, 128);
	log_info("new virtio device on '%s' (vid=%d)", ifname, vid);
	log_debug("Number of queue pairs=%u\n", num_vring_pairs);

	/* disable notifications. */
	for (unsigned q=0; q<num_vring_pairs; ++q) {
		rte_vhost_enable_guest_notification(vid, q*2, 0); /* RX queue */
		rte_vhost_enable_guest_notification(vid, q*2+1, 0); /* TX queue */
	}

	id = get_relay_for_sock(ifname);
	if (id < 0) {
		log_error("Could not find relay for %s", ifname);
		return -1;
	}

	if (virtio_forwarder_add_virtio(vid, id) == 0)
		return 0;
	else
		log_warning("Could not add virtio worker for '%s'", ifname);

	return -1;
}

static void virtio_vhostuser_destroy_device_cb(int vid)
{
	char str[128];
	int id;

	rte_vhost_get_ifname(vid, str, 128);
	log_info("destroy virtio device on '%s'", str);
	id = get_relay_for_sock(str);
	if (id < 0) {
		log_error("Invalid relay id found for vhostuser fd '%s'", str);
		return;
	}
	virtio_forwarder_remove_virtio(id);
}

static int
virtio_vhostuser_vring_state_change_cb(int vid, uint16_t queue_id, int enable)
{
	char ifname[128];
	int id;

	rte_vhost_get_ifname(vid, ifname, 128);
	id = get_relay_for_sock(ifname);
	if (id < 0) {
		log_error("Invalid relay id found for vhostuser fd '%s'", ifname);
		return -1;
	}
	log_debug("vring state change on queue_id=%hu (enable=%d) on relay %d ('%s')",
		queue_id, enable, id, ifname);
	virtio_forwarder_vring_state_change(id, queue_id, enable);

	return 0;
}

#if RTE_VERSION >= RTE_VERSION_NUM(17,5,0,0)
static int virtio_vhostuser_features_changed_cb(int vid, uint64_t features)
{
	char ifname[128];
	int id;

	rte_vhost_get_ifname(vid, ifname, 128);
	id = get_relay_for_sock(ifname);
	if (id < 0) {
		log_error("Invalid relay id found for vhostuser fd '%s'", ifname);
		return -1;
	}
	log_debug("Feature change on relay %d ('%s'). New features=%lu", id,
		ifname, features);

	/* From the docs:
	 * >VHOST_F_LOG_ALL will be set/cleared at the start/end of live
	 * >migration, respectively.
	 *
	 * RTE_VHOST_NEED_LOG() checks for VHOST_F_LOG_ALL.
	 * This check ensures that we do not call rte_vhost_avail_entries
	 * (virtio_worker.c, dpdk_rx) too late during the migration procedure. A
	 * segmentation fault results if we do: Note that this may actually be a
	 * DPDK issue, but for now we workaround it using the lm_pending flag.
	 */
	if (RTE_VHOST_NEED_LOG(features)) {
		log_info("VHOST_F_LOG_ALL is set. Live migration is being initiated on virtio %d",
			id);
		virtio_set_lm_pending(id);
	}

	return 0;
}
#endif

#if RTE_VERSION >= RTE_VERSION_NUM(17,11,0,0)
static int virtio_vhostuser_new_connection_cb(int vid)
{
	char ifname[128];

	rte_vhost_get_ifname(vid, ifname, 128);
	log_debug("new virtio connection on '%s' (vid=%d)", ifname, vid);

	return 0;
}

static void virtio_vhostuser_destroy_connection_cb(int vid)
{
	char ifname[128];

	rte_vhost_get_ifname(vid, ifname, 128);
	log_debug("destroy virtio connection on '%s'", ifname);

	return;
}
#endif
#else
static int virtio_vhostuser_new_device_cb(struct virtio_net *dev)
{
	int id;

	log_info("new virtio device on '%s': features=0x%08llX,protocol_features=0x%08llX,virt_qp_nb=%u,device_fh=%"PRIu64",q0_len=%u,q1_len=%u",
		dev->ifname, (long long)dev->features,
		(long long)dev->protocol_features, dev->virt_qp_nb,
		dev->device_fh, dev->virtqueue[0]->size, dev->virtqueue[1]->size);

	if (get_log_level() >= LOG_LVL_DEBUG) {
		char buf[512];
		unsigned i;
		unsigned l=0;
		buf[0]=0;
		for (i=0; i<=(sizeof(virtio_features) / sizeof(*virtio_features)); ++i) {
			if (dev->features & (1ULL<<i))
				l+=snprintf(buf+l, 512-l, "%s,", virtio_features[i]);
		}
		if (l) {
			buf[l-1]=0;
			log_debug("VirtIO features: %s", buf);
		}
		buf[0]=0;
		l=0;
		for (i=0; i<=(sizeof(vhost_features) / sizeof(*vhost_features)); ++i) {
			if (dev->protocol_features & (1ULL<<i))
				l+=snprintf(buf+l, 512-l, "%s,", vhost_features[i]);
		}
		if (l) {
			buf[l-1]=0;
			log_debug("Vhost-user features: %s", buf);
		}
		for (i=0; i<=1; ++i)
			log_debug("virtqueue %u: size=%u, last_used_idx=%hu, avail_idx=%hu, used_idx=%hu",
				i, dev->virtqueue[i]->size,
				dev->virtqueue[i]->last_used_idx,
				dev->virtqueue[i]->avail->idx,
				dev->virtqueue[i]->used->idx);
	}
	// disable notifications
	for (unsigned q=0; q<dev->virt_qp_nb; ++q) {
		rte_vhost_enable_guest_notification(dev, q*2, 0); // RX queue
		rte_vhost_enable_guest_notification(dev, q*2+1, 0); // TX queue
	}

	if ((id = virtio_vhostuser_name_to_id(dev->ifname)) >= 0
			&& virtio_forwarder_add_virtio(dev, id) == 0) {
		dev->flags |= VIRTIO_DEV_RUNNING;
		return 0;
	}
	log_warning("Could not add virtio worker for '%s'", dev->ifname);

	return -1;
}

static void virtio_vhostuser_destroy_device_cb(volatile struct virtio_net *dev)
{
	char str[128];
	int i=0;
	while (i<127) {
		str[i]=dev->ifname[i];
		if (str[i]==0)
			break;
		++i;
	}
	str[127]=0;
	int id = virtio_vhostuser_name_to_id(str);
	log_info("destroy virtio device on '%s'", str);
	dev->flags &= ~VIRTIO_DEV_RUNNING;
	if (id<0) {
		log_error("Invalid id found in vhostuser fd name '%s'", dev->ifname);
		return;
	}
	virtio_forwarder_remove_virtio(id);
}

static int
virtio_vhostuser_vring_state_change_cb(struct virtio_net *dev,
			uint16_t queue_id, int enable)
{
	char str[128];
	int i=0;
	while (i<127) {
		str[i]=dev->ifname[i];
		if (str[i]==0)
			break;
		++i;
	}
	str[127]=0;
	int id = virtio_vhostuser_name_to_id(str);
	log_debug("vring state change on queue_id=%hu (enable=%d) on '%s'", queue_id, enable, str);
	if (id<0) {
		log_error("Invalid id found in vhostuser fd name '%s'", str);
		return -1;
	}
	virtio_forwarder_vring_state_change(id, queue_id, enable);
	return 0;
}
#endif

#if RTE_VERSION >= RTE_VERSION_NUM(17,5,0,0)
static const struct vhost_device_ops virtio_vhostuser_ops = {
	.features_changed = virtio_vhostuser_features_changed_cb,
#else
static const struct virtio_net_device_ops virtio_vhostuser_ops = {
#endif
	.new_device =  virtio_vhostuser_new_device_cb,
	.destroy_device = virtio_vhostuser_destroy_device_cb,
	.vring_state_changed = virtio_vhostuser_vring_state_change_cb,
#if RTE_VERSION >= RTE_VERSION_NUM(17,11,0,0)
	.new_connection = virtio_vhostuser_new_connection_cb,
	.destroy_connection = virtio_vhostuser_destroy_connection_cb,
#endif
};

#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
static void* virtio_vhostuser_threadmain(void *arg __attribute__((unused)))
{
	log_debug("Starting vhostuser thread");
	rte_vhost_driver_session_start();
	log_debug("vhostuser thread ending");

	return 0;
}
#endif

static int prep_default_sockets(void)
{
	struct stat fstat;
	bool created_path = false;
	char *s;
	int len;
	uid_t uid, gid;
	const struct virtio_vhostuser_conf *conf = &g_vio_worker_conf;

	/* Check if the provided directory exists and attempt autcreate if not. */
	if (stat(conf->socket_path, &fstat) == 0) {
		if (!S_ISDIR(fstat.st_mode)) {
			log_error("Provided vhost-user path ('%s') is not a valid path",
				conf->socket_path);
			return 1;
		}
	} else {
		// try to autocreate
		log_warning("Could not access provided vhost-user socket path '%s' (%m), trying to create",
			conf->socket_path);
		if (mkdir(conf->socket_path, 0777)!=0) {
			log_error("Error creating provided vhost-user socket path '%s': %m",
			conf->socket_path);
			return 1;
		}
		created_path = true;
	}

	/* Check if the template name is valid. */
	if ((s=strstr(conf->socket_name, "%u")) == 0) {
		log_error("Mandatory single '%%u' not found in provided vhost-user socket name '%s'",
			conf->socket_name);
		return 1;
	} else {
		if (strstr(s+1, "%u") != 0) {
			log_error("More than one occurrence of '%%u' found in provided vhost-user socket name '%s'",
				conf->socket_name);
			return 1;
		}
	}
	if (strchr(conf->socket_name, '/')) {
		log_error("Invalid '/' found in vhost-user socket name '%s'",
			conf->socket_name);
		return 1;
	}

	/* Name seems valid. */
	strncpy(vhost_socket_name_suffix, conf->socket_name, 128); /* temp copy of const conf string */
	s = strstr(vhost_socket_name_suffix, "%u");
	len = snprintf(vhost_socket_name_prefix, 128, "%s/", conf->socket_path);
	*s=0;
	snprintf(vhost_socket_name_prefix+len, 128-len, "%s", vhost_socket_name_suffix); /* append up to %u */
	if (*(s+2)) {// if something more after %u
		strncpy(vhost_socket_name_suffix, s+2, 128);
		if (vhost_socket_name_suffix[0] >= '0' &&
				vhost_socket_name_suffix[0] <= '9') {
			log_error("vhost-user socket name '%s' may not contain a number after '%%u'!'",
				conf->socket_name);
			return 1;
		}
	} else
		vhost_socket_name_suffix[0] = 0;

	/* Change ownership of vhost folder if we created it. */
	uid = getuid();
	gid = getgid();
	if (conf->vhost_username[0] && get_uid(conf->vhost_username, &uid) != 0) {
		log_error("Could not get uid for '%s'!", conf->vhost_username);
		return -1;
	}
	if (conf->vhost_groupname[0] && get_gid(conf->vhost_groupname, &gid) != 0) {
		log_error("Could not get gid for '%s'!", conf->vhost_groupname);
		return -1;
	}
	if ((uid != getuid() || gid != getgid()) && created_path) {
		if (chown(conf->socket_path, uid, gid) != 0) {
			log_warning("Could not set ownership of socket path '%s'!",
				conf->socket_path);
		}
	}
	log_debug("Using vhost-user prefix '%s' and suffix '%s', with uid=%u and gid=%u",
		vhost_socket_name_prefix, vhost_socket_name_suffix, uid, gid);

	return 0;
}

static int register_relay_socket(const char *vhost_path, unsigned relay_id)
{
	struct stat statbuf;
#if RTE_VERSION >= RTE_VERSION_NUM(16,7,0,0)
	unsigned flags = 0;
#endif
	uid_t uid, gid;
	const struct virtio_vhostuser_conf *conf = &g_vio_worker_conf;

	log_debug("Got register_relay_socket(%s, %u)", vhost_path, relay_id);

	/* Check for pre-existing sockets if server. */
	if (conf->vhost_client == 0 && stat(vhost_path, &statbuf) == 0) {
		/* virtio-forwarder should not get this far if the
		 * daemon is already running (see pidfile check in
		 * virtio_forwarder_main.c). */
		if (S_ISSOCK(statbuf.st_mode)) {
			if (unlink(vhost_path) != 0) {
				log_warning("Found stale unix socket handle '%s', could not delete!",
					vhost_path);
				return -1;
			}
			log_warning("Deleted stale unix socket handle '%s'",
				vhost_path);
		} else {
			log_error("'%s' unexpectedly already exists!", vhost_path);
			return -1;
		}
	}

	if (relay_ifname_map[relay_id]) {
		log_error("Cannot register %s! Socket %s is already assigned to relay %u",
			vhost_path, relay_ifname_map[relay_id], relay_id);
		return -1;
	}

#if RTE_VERSION >= RTE_VERSION_NUM(16,7,0,0)
	if (conf->vhost_client)
		flags |= RTE_VHOST_USER_CLIENT;
#endif
#if RTE_VERSION >= RTE_VERSION_NUM(16,11,0,0)
	if (conf->zerocopy)
		flags |= RTE_VHOST_USER_DEQUEUE_ZERO_COPY;
#endif
#if RTE_VERSION >= RTE_VERSION_NUM(16,7,0,0)
	if (rte_vhost_driver_register(vhost_path, flags) != 0) {
#else
	if (rte_vhost_driver_register(vhost_path) != 0) {
#endif
		log_error("Error starting vhost user instance '%s'!", vhost_path);
		return -1;
	}

#if RTE_VERSION >= RTE_VERSION_NUM(17,5,0,0)
	/* Set features. */
	if (conf->use_rx_mrgbuf == 0) {
		if (rte_vhost_driver_disable_features(
				vhost_path, 1ULL << VIRTIO_NET_F_MRG_RXBUF))
			log_warning("Failed to disabe mergeable RX buffers on %s",
				vhost_path);
	}
	if (rte_vhost_driver_disable_features(
			vhost_path, 1ULL << VIRTIO_NET_F_CTRL_RX))
		log_warning("Failed to disable RX control commands on %s",
			vhost_path);
	if (conf->enable_tso == 0) {
		int disable_tso = 0;
		disable_tso |= rte_vhost_driver_disable_features(
					vhost_path, 1ULL << VIRTIO_NET_F_HOST_TSO4);
		disable_tso |= rte_vhost_driver_disable_features(
					vhost_path, 1ULL << VIRTIO_NET_F_HOST_TSO6);
		if (disable_tso)
			log_warning("Failed to disable TSO on %s", vhost_path);
	}

	/* Start vhost diver for socket. */
	rte_vhost_driver_callback_register(vhost_path, &virtio_vhostuser_ops);
	if (rte_vhost_driver_start(vhost_path)) {
		log_critical("Error starting vhost driver on socket %s!",
			vhost_path);
		return -1;
	}
#endif
	relay_ifname_map[relay_id] = (char *)malloc(strlen(vhost_path)+1);
	strcpy(relay_ifname_map[relay_id], vhost_path);

	/* Change ownership of the socket files. */
	uid = getuid();
	gid = getgid();
	get_uid(conf->vhost_username, &uid);
	get_gid(conf->vhost_groupname, &gid);
	if (conf->vhost_client == 0 && (uid != getuid() || gid != getgid())) {
		if (chown(vhost_path, uid, gid) != 0) {
			log_error("Could not change ownership of '%s' (%m)!",
				vhost_path);
			return -1;
		}
	}

	return 0;
}

static int deregister_socket(const char *vhost_path)
{
	int err, id;

	log_debug("Got deregister_socket(%s)", vhost_path);

	err = rte_vhost_driver_unregister(vhost_path);
	if (err != 0)
		log_warning("rte_vhost_driver_unregister(%s) failed with error %d",
			vhost_path, err);

	id = get_relay_for_sock(vhost_path);
	if (id < 0) {
		log_error("Found invalid relay id=%d during %s deregistration",
			id, vhost_path);
		return 1;
	}
	if (relay_ifname_map[id]) {
		free(relay_ifname_map[id]);
		relay_ifname_map[id] = NULL;
	} else {
		log_error("Encountered unexpected NULL pointer relay_ifname_map[%d]",
			id);
		return 2;
	}

	return err;
}

int virtio_vhostuser_start(const struct virtio_vhostuser_conf *conf,
			bool create_sockets)
{
	int err;

	mk_default_sockets = create_sockets;
#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
	/* TODO: Investigate dynamic socket support for old DPDKs. */
	if (!mk_default_sockets) {
		log_warning("Dynamic sockets not supported for this version of DPDK!");
		mk_default_sockets = true;
	}
#endif

	memset(relay_ifname_map, 0, sizeof(relay_ifname_map));
	g_vio_worker_conf = *conf;

	if (mk_default_sockets) {
		log_debug("Creating default sockets for %d relays", MAX_RELAYS);
		err = prep_default_sockets();
		if (err)
			return err;
	} else {
		log_debug("Using dynamic socket connection API");
	}

	log_debug("Initializing virtio workers");
	if (virtio_forwarders_initialize() != 0) {
		log_error("Error starting worker threads!");
		return -1;
	}

	log_debug("Registering vhost driver");
#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
	if (conf->use_rx_mrgbuf == 0)
		/* Enabling RX buffer merging can cause 50% performance
		 * degradation with DPDK <=16.07, and 20% degradation with DPDK 16.11! */
		rte_vhost_feature_disable((1ULL << VIRTIO_NET_F_MRG_RXBUF));

	/* VIRTIO_NET_F_CTRL_RX Requires VIRTIO_NET_F_CTRL_VQ. */
	rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_CTRL_RX);
	if (conf->enable_tso == 0) {
		rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_HOST_TSO4);
		rte_vhost_feature_disable(1ULL << VIRTIO_NET_F_HOST_TSO6);
	}
#endif

	if (mk_default_sockets) {
		for (int id=0; id<MAX_RELAYS; ++id) {
			char buf[128];

			virtio_vhostuser_id_to_name(id, buf, 128);
			err = register_relay_socket(buf, id);
			if (err)
				return err;

		}
	}

#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
	rte_vhost_driver_callback_register(&virtio_vhostuser_ops);
	log_debug("Creating vhostuser thread");
	pthread_create(&vhostuser_thread, 0, virtio_vhostuser_threadmain, 0);
#endif
	return 0;
}

void virtio_vhostuser_stop(void) {
	virtio_forwarders_remove_all();
	log_debug("Stopping vhostuser thread");
	for (int id=0; id<MAX_RELAYS; ++id) {
		if (relay_ifname_map[id])
			deregister_socket(relay_ifname_map[id]);
	}

#if RTE_VERSION < RTE_VERSION_NUM(17,5,0,0)
	struct timespec ts;
	struct timeval tv;
	gettimeofday(&tv, 0);
	tv.tv_sec+=1;
	ts.tv_sec=tv.tv_sec;
	ts.tv_nsec=tv.tv_usec*1000;
	if (pthread_timedjoin_np(vhostuser_thread, 0, &ts) != 0) {
		log_debug("Timeout waiting for vhostuser_thread, cancelling thread...");
		pthread_cancel(vhostuser_thread);
	}
	pthread_join(vhostuser_thread, 0);
#endif
	virtio_forwarders_shutdown();
	log_debug("Stopped vhostuser thread");
}

/* API for dynamic socket operations. */
int virtio_add_sock_dev_pair(const char *vhost_path,
			char slave_dbdfs[MAX_NUM_BOND_SLAVES][RTE_ETH_NAME_MAX_LEN],
			unsigned num_slaves, char *name, uint8_t mode,
			bool conditional)
{
	int err;
	int relay_id;
	char *dev;
#if RTE_VERSION >= RTE_VERSION_NUM(16,7,0,0)
	dpdk_port_t port_id;
#endif

	dev = num_slaves == 1 ? slave_dbdfs[0] : name;
#if RTE_VERSION >= RTE_VERSION_NUM(16,7,0,0)
	if (rte_eth_dev_get_port_by_name(dev, &port_id) == 0) {
		log_error("Device %s is already attached to virtio-forwarder",
			dev);
		return -EEXIST;
	}
#endif

	for (int i=0; i<MAX_RELAYS; ++i) {
		if (relay_ifname_map[i]) {
			if (strcmp(relay_ifname_map[i], vhost_path) == 0) {
				log_error("Socket %s is already registered with relay %d",
					vhost_path, i);
				return 1;
			}
		}
	}

	relay_id = virtio_get_free_relay_id(relay_ifname_map);
	if (relay_id < 0) {
		log_error("Could not get idle relay for packet processing!");
		return 2;
	}

	/* Register VF or bond. */
	if (num_slaves == 1) {
		err = virtio_forwarder_add_vf2(dev, relay_id, conditional);
		if (err != 0) {
			log_error("virtio_forwarder_add_vf2(%s, %d, %s) failed with error %d",
				dev, relay_id, conditional ? "true" : "false",
				err);
			return 3;
		}
	} else {
		err = virtio_forwarder_bond_add(slave_dbdfs, num_slaves, name,
						mode, relay_id);
		if (err != 0) {
			log_error("virtio_forwarder_bond_add(<PCI addresses>, %u, %s, %u, %d) failed with error %d",
				num_slaves, name, mode, relay_id, err);
			return 4;
		}
	}

	/* Register socket. */
	err = register_relay_socket(vhost_path, relay_id);
	if (err != 0) {
		log_error("register_relay_socket failed with error %d", err);
		goto exit_deconfigure_pair;
	}

	return err;

exit_deconfigure_pair:
	err = virtio_forwarder_remove_vf2(dev, relay_id, conditional);
	if (err)
		log_warning("During error recovery, virtio_forwarder_remove_vf2 failed with error %d",
			err);

	err = deregister_socket(vhost_path);
	if (err)
		log_warning("During error recovery, deregister_socket failed with error %d",
			err);

	return 5;
}

int virtio_remove_sock_dev_pair(const char *vhost_path, const char *dev,
			bool conditional)
{
	int err;
	int relay_id;

	log_debug("Got virtio_remove_sock_dev_pair(%s, %s, %s)", vhost_path,
		dev, conditional ? "true" : "false");

	relay_id = get_relay_for_sock(vhost_path);
	if (relay_id < 0) {
		log_error("Could not get relay id for vhostuser fd %s!",
			vhost_path);
		return 1;
	}

	/* Check that the device <-> relay pairing is correct. */
	if (!virtio_relay_has_device(relay_id, dev)) {
		log_error("The relay instance for socket %s has no device named %s. Not removing pair",
			vhost_path, dev);
		return 2;
	}

	/* Remove virtio: VM may or may not have shutdown at this point.
	 * Therefore, remove virtio manually to cater for case where the device
	 * destroy callback did not trigger yet. */
	virtio_forwarder_remove_virtio(relay_id);

	/* Deregister socket. */
	err = deregister_socket(vhost_path);
	if (err != 0) {
		log_error("deregister_socket(%s) failed with error %d",
			vhost_path, err);
		return 3;
	}

	/* Remove VF. */
	err = virtio_forwarder_remove_vf2(dev, relay_id, conditional);
	if (err != 0) {
		log_error("virtio_forwarder_remove_vf2(%s, %d, %s) failed with error %d",
			dev, relay_id, conditional ? "true" : "false",  err);
		return 4;
	}

	return err;
}