summaryrefslogtreecommitdiffstats
path: root/src/spdk/lib/virtio/virtio.c
blob: b03034cffc8df2d3fb4290d2007429a3960bd5de (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
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
/*-
 *   BSD LICENSE
 *
 *   Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
 *   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 Intel Corporation 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 "spdk/stdinc.h"

#include <linux/virtio_scsi.h>
#include <linux/virtio_pci.h>
#include <linux/virtio_config.h>

#include <rte_config.h>
#include <rte_memcpy.h>
#include <rte_string_fns.h>
#include <rte_memzone.h>
#include <rte_malloc.h>
#include <rte_atomic.h>
#include <rte_branch_prediction.h>
#include <rte_pci.h>
#include <rte_common.h>
#include <rte_errno.h>

#include <rte_eal.h>
#include <rte_dev.h>
#include <rte_prefetch.h>

#include "spdk/env.h"
#include "spdk/barrier.h"

#include "spdk_internal/virtio.h"

/* We use SMP memory barrier variants as all virtio_pci devices
 * are purely virtual. All MMIO is executed on a CPU core, so
 * there's no need to do full MMIO synchronization.
 */
#define virtio_mb()	spdk_smp_mb()
#define virtio_rmb()	spdk_smp_rmb()
#define virtio_wmb()	spdk_smp_wmb()

/* Chain all the descriptors in the ring with an END */
static inline void
vring_desc_init(struct vring_desc *dp, uint16_t n)
{
	uint16_t i;

	for (i = 0; i < n - 1; i++) {
		dp[i].next = (uint16_t)(i + 1);
	}
	dp[i].next = VQ_RING_DESC_CHAIN_END;
}

static void
virtio_init_vring(struct virtqueue *vq)
{
	int size = vq->vq_nentries;
	struct vring *vr = &vq->vq_ring;
	uint8_t *ring_mem = vq->vq_ring_virt_mem;

	/*
	 * Reinitialise since virtio port might have been stopped and restarted
	 */
	memset(ring_mem, 0, vq->vq_ring_size);
	vring_init(vr, size, ring_mem, VIRTIO_PCI_VRING_ALIGN);
	vq->vq_used_cons_idx = 0;
	vq->vq_desc_head_idx = 0;
	vq->vq_avail_idx = 0;
	vq->vq_desc_tail_idx = (uint16_t)(vq->vq_nentries - 1);
	vq->vq_free_cnt = vq->vq_nentries;
	vq->req_start = VQ_RING_DESC_CHAIN_END;
	vq->req_end = VQ_RING_DESC_CHAIN_END;
	vq->reqs_finished = 0;
	memset(vq->vq_descx, 0, sizeof(struct vq_desc_extra) * vq->vq_nentries);

	vring_desc_init(vr->desc, size);

	/* Tell the backend not to interrupt us.
	 * If F_EVENT_IDX is negotiated, we will always set incredibly high
	 * used event idx, so that we will practically never receive an
	 * interrupt. See virtqueue_req_flush()
	 */
	if (vq->vdev->negotiated_features & (1ULL << VIRTIO_RING_F_EVENT_IDX)) {
		vring_used_event(&vq->vq_ring) = UINT16_MAX;
	} else {
		vq->vq_ring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT;
	}
}

static int
virtio_init_queue(struct virtio_dev *dev, uint16_t vtpci_queue_idx)
{
	unsigned int vq_size, size;
	struct virtqueue *vq;
	int rc;

	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "setting up queue: %"PRIu16"\n", vtpci_queue_idx);

	/*
	 * Read the virtqueue size from the Queue Size field
	 * Always power of 2 and if 0 virtqueue does not exist
	 */
	vq_size = virtio_dev_backend_ops(dev)->get_queue_size(dev, vtpci_queue_idx);
	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "vq_size: %u\n", vq_size);
	if (vq_size == 0) {
		SPDK_ERRLOG("virtqueue %"PRIu16" does not exist\n", vtpci_queue_idx);
		return -EINVAL;
	}

	if (!rte_is_power_of_2(vq_size)) {
		SPDK_ERRLOG("virtqueue %"PRIu16" size (%u) is not powerof 2\n",
			    vtpci_queue_idx, vq_size);
		return -EINVAL;
	}

	size = RTE_ALIGN_CEIL(sizeof(*vq) +
			      vq_size * sizeof(struct vq_desc_extra),
			      RTE_CACHE_LINE_SIZE);

	vq = spdk_dma_zmalloc(size, RTE_CACHE_LINE_SIZE, NULL);
	if (vq == NULL) {
		SPDK_ERRLOG("can not allocate vq\n");
		return -ENOMEM;
	}
	dev->vqs[vtpci_queue_idx] = vq;

	vq->vdev = dev;
	vq->vq_queue_index = vtpci_queue_idx;
	vq->vq_nentries = vq_size;

	/*
	 * Reserve a memzone for vring elements
	 */
	size = vring_size(vq_size, VIRTIO_PCI_VRING_ALIGN);
	vq->vq_ring_size = RTE_ALIGN_CEIL(size, VIRTIO_PCI_VRING_ALIGN);
	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "vring_size: %u, rounded_vring_size: %u\n",
		      size, vq->vq_ring_size);

	vq->owner_thread = NULL;

	rc = virtio_dev_backend_ops(dev)->setup_queue(dev, vq);
	if (rc < 0) {
		SPDK_ERRLOG("setup_queue failed\n");
		spdk_dma_free(vq);
		dev->vqs[vtpci_queue_idx] = NULL;
		return rc;
	}

	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "vq->vq_ring_mem:      0x%" PRIx64 "\n",
		      vq->vq_ring_mem);
	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "vq->vq_ring_virt_mem: 0x%" PRIx64 "\n",
		      (uint64_t)(uintptr_t)vq->vq_ring_virt_mem);

	virtio_init_vring(vq);
	return 0;
}

static void
virtio_free_queues(struct virtio_dev *dev)
{
	uint16_t nr_vq = dev->max_queues;
	struct virtqueue *vq;
	uint16_t i;

	if (dev->vqs == NULL) {
		return;
	}

	for (i = 0; i < nr_vq; i++) {
		vq = dev->vqs[i];
		if (!vq) {
			continue;
		}

		virtio_dev_backend_ops(dev)->del_queue(dev, vq);

		rte_free(vq);
		dev->vqs[i] = NULL;
	}

	rte_free(dev->vqs);
	dev->vqs = NULL;
}

static int
virtio_alloc_queues(struct virtio_dev *dev, uint16_t request_vq_num, uint16_t fixed_vq_num)
{
	uint16_t nr_vq;
	uint16_t i;
	int ret;

	nr_vq = request_vq_num + fixed_vq_num;
	if (nr_vq == 0) {
		/* perfectly fine to have a device with no virtqueues. */
		return 0;
	}

	assert(dev->vqs == NULL);
	dev->vqs = rte_zmalloc(NULL, sizeof(struct virtqueue *) * nr_vq, 0);
	if (!dev->vqs) {
		SPDK_ERRLOG("failed to allocate %"PRIu16" vqs\n", nr_vq);
		return -ENOMEM;
	}

	for (i = 0; i < nr_vq; i++) {
		ret = virtio_init_queue(dev, i);
		if (ret < 0) {
			virtio_free_queues(dev);
			return ret;
		}
	}

	dev->max_queues = nr_vq;
	dev->fixed_queues_num = fixed_vq_num;
	return 0;
}

/**
 * Negotiate virtio features. For virtio_user this will also set
 * dev->modern flag if VIRTIO_F_VERSION_1 flag is negotiated.
 */
static int
virtio_negotiate_features(struct virtio_dev *dev, uint64_t req_features)
{
	uint64_t host_features = virtio_dev_backend_ops(dev)->get_features(dev);
	int rc;

	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "guest features = %" PRIx64 "\n", req_features);
	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "device features = %" PRIx64 "\n", host_features);

	rc = virtio_dev_backend_ops(dev)->set_features(dev, req_features & host_features);
	if (rc != 0) {
		SPDK_ERRLOG("failed to negotiate device features.\n");
		return rc;
	}

	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "negotiated features = %" PRIx64 "\n",
		      dev->negotiated_features);

	virtio_dev_set_status(dev, VIRTIO_CONFIG_S_FEATURES_OK);
	if (!(virtio_dev_get_status(dev) & VIRTIO_CONFIG_S_FEATURES_OK)) {
		SPDK_ERRLOG("failed to set FEATURES_OK status!\n");
		/* either the device failed, or we offered some features that
		 * depend on other, not offered features.
		 */
		return -EINVAL;
	}

	return 0;
}

int
virtio_dev_construct(struct virtio_dev *vdev, const char *name,
		     const struct virtio_dev_ops *ops, void *ctx)
{
	int rc;

	vdev->name = strdup(name);
	if (vdev->name == NULL) {
		return -ENOMEM;
	}

	rc = pthread_mutex_init(&vdev->mutex, NULL);
	if (rc != 0) {
		free(vdev->name);
		return -rc;
	}

	vdev->backend_ops = ops;
	vdev->ctx = ctx;

	return 0;
}

int
virtio_dev_reset(struct virtio_dev *dev, uint64_t req_features)
{
	req_features |= (1ULL << VIRTIO_F_VERSION_1);

	virtio_dev_stop(dev);

	virtio_dev_set_status(dev, VIRTIO_CONFIG_S_ACKNOWLEDGE);
	if (!(virtio_dev_get_status(dev) & VIRTIO_CONFIG_S_ACKNOWLEDGE)) {
		SPDK_ERRLOG("Failed to set VIRTIO_CONFIG_S_ACKNOWLEDGE status.\n");
		return -EIO;
	}

	virtio_dev_set_status(dev, VIRTIO_CONFIG_S_DRIVER);
	if (!(virtio_dev_get_status(dev) & VIRTIO_CONFIG_S_DRIVER)) {
		SPDK_ERRLOG("Failed to set VIRTIO_CONFIG_S_DRIVER status.\n");
		return -EIO;
	}

	return virtio_negotiate_features(dev, req_features);
}

int
virtio_dev_start(struct virtio_dev *vdev, uint16_t max_queues, uint16_t fixed_queue_num)
{
	int ret;

	ret = virtio_alloc_queues(vdev, max_queues, fixed_queue_num);
	if (ret < 0) {
		return ret;
	}

	virtio_dev_set_status(vdev, VIRTIO_CONFIG_S_DRIVER_OK);
	if (!(virtio_dev_get_status(vdev) & VIRTIO_CONFIG_S_DRIVER_OK)) {
		SPDK_ERRLOG("Failed to set VIRTIO_CONFIG_S_DRIVER_OK status.\n");
		return -1;
	}

	return 0;
}

void
virtio_dev_destruct(struct virtio_dev *dev)
{
	virtio_dev_backend_ops(dev)->destruct_dev(dev);
	pthread_mutex_destroy(&dev->mutex);
	free(dev->name);
}

static void
vq_ring_free_chain(struct virtqueue *vq, uint16_t desc_idx)
{
	struct vring_desc *dp, *dp_tail;
	struct vq_desc_extra *dxp;
	uint16_t desc_idx_last = desc_idx;

	dp  = &vq->vq_ring.desc[desc_idx];
	dxp = &vq->vq_descx[desc_idx];
	vq->vq_free_cnt = (uint16_t)(vq->vq_free_cnt + dxp->ndescs);
	if ((dp->flags & VRING_DESC_F_INDIRECT) == 0) {
		while (dp->flags & VRING_DESC_F_NEXT) {
			desc_idx_last = dp->next;
			dp = &vq->vq_ring.desc[dp->next];
		}
	}
	dxp->ndescs = 0;

	/*
	 * We must append the existing free chain, if any, to the end of
	 * newly freed chain. If the virtqueue was completely used, then
	 * head would be VQ_RING_DESC_CHAIN_END (ASSERTed above).
	 */
	if (vq->vq_desc_tail_idx == VQ_RING_DESC_CHAIN_END) {
		vq->vq_desc_head_idx = desc_idx;
	} else {
		dp_tail = &vq->vq_ring.desc[vq->vq_desc_tail_idx];
		dp_tail->next = desc_idx;
	}

	vq->vq_desc_tail_idx = desc_idx_last;
	dp->next = VQ_RING_DESC_CHAIN_END;
}

static uint16_t
virtqueue_dequeue_burst_rx(struct virtqueue *vq, void **rx_pkts,
			   uint32_t *len, uint16_t num)
{
	struct vring_used_elem *uep;
	struct virtio_req *cookie;
	uint16_t used_idx, desc_idx;
	uint16_t i;

	/*  Caller does the check */
	for (i = 0; i < num ; i++) {
		used_idx = (uint16_t)(vq->vq_used_cons_idx & (vq->vq_nentries - 1));
		uep = &vq->vq_ring.used->ring[used_idx];
		desc_idx = (uint16_t) uep->id;
		len[i] = uep->len;
		cookie = (struct virtio_req *)vq->vq_descx[desc_idx].cookie;

		if (spdk_unlikely(cookie == NULL)) {
			SPDK_WARNLOG("vring descriptor with no mbuf cookie at %"PRIu16"\n",
				     vq->vq_used_cons_idx);
			break;
		}

		rte_prefetch0(cookie);
		rx_pkts[i]  = cookie;
		vq->vq_used_cons_idx++;
		vq_ring_free_chain(vq, desc_idx);
		vq->vq_descx[desc_idx].cookie = NULL;
	}

	return i;
}

static void
finish_req(struct virtqueue *vq)
{
	struct vring_desc *desc;
	uint16_t avail_idx;

	desc = &vq->vq_ring.desc[vq->req_end];
	desc->flags &= ~VRING_DESC_F_NEXT;

	/*
	 * Place the head of the descriptor chain into the next slot and make
	 * it usable to the host. The chain is made available now rather than
	 * deferring to virtqueue_req_flush() in the hopes that if the host is
	 * currently running on another CPU, we can keep it processing the new
	 * descriptor.
	 */
	avail_idx = (uint16_t)(vq->vq_avail_idx & (vq->vq_nentries - 1));
	vq->vq_ring.avail->ring[avail_idx] = vq->req_start;
	vq->vq_avail_idx++;
	vq->req_end = VQ_RING_DESC_CHAIN_END;
	virtio_wmb();
	vq->vq_ring.avail->idx = vq->vq_avail_idx;
	vq->reqs_finished++;
}

int
virtqueue_req_start(struct virtqueue *vq, void *cookie, int iovcnt)
{
	struct vq_desc_extra *dxp;

	if (iovcnt > vq->vq_free_cnt) {
		return iovcnt > vq->vq_nentries ? -EINVAL : -ENOMEM;
	}

	if (vq->req_end != VQ_RING_DESC_CHAIN_END) {
		finish_req(vq);
	}

	vq->req_start = vq->vq_desc_head_idx;
	dxp = &vq->vq_descx[vq->req_start];
	dxp->cookie = cookie;
	dxp->ndescs = 0;

	return 0;
}

void
virtqueue_req_flush(struct virtqueue *vq)
{
	uint16_t reqs_finished;

	if (vq->req_end == VQ_RING_DESC_CHAIN_END) {
		/* no non-empty requests have been started */
		return;
	}

	finish_req(vq);
	virtio_mb();

	reqs_finished = vq->reqs_finished;
	vq->reqs_finished = 0;

	if (vq->vdev->negotiated_features & (1ULL << VIRTIO_RING_F_EVENT_IDX)) {
		/* Set used event idx to a value the device will never reach.
		 * This effectively disables interrupts.
		 */
		vring_used_event(&vq->vq_ring) = vq->vq_used_cons_idx - vq->vq_nentries - 1;

		if (!vring_need_event(vring_avail_event(&vq->vq_ring),
				      vq->vq_avail_idx,
				      vq->vq_avail_idx - reqs_finished)) {
			return;
		}
	} else if (vq->vq_ring.used->flags & VRING_USED_F_NO_NOTIFY) {
		return;
	}

	virtio_dev_backend_ops(vq->vdev)->notify_queue(vq->vdev, vq);
	SPDK_DEBUGLOG(SPDK_LOG_VIRTIO_DEV, "Notified backend after xmit\n");
}

void
virtqueue_req_abort(struct virtqueue *vq)
{
	struct vring_desc *desc;

	if (vq->req_start == VQ_RING_DESC_CHAIN_END) {
		/* no requests have been started */
		return;
	}

	desc = &vq->vq_ring.desc[vq->req_end];
	desc->flags &= ~VRING_DESC_F_NEXT;

	vq_ring_free_chain(vq, vq->req_start);
	vq->req_start = VQ_RING_DESC_CHAIN_END;
}

void
virtqueue_req_add_iovs(struct virtqueue *vq, struct iovec *iovs, uint16_t iovcnt,
		       enum spdk_virtio_desc_type desc_type)
{
	struct vring_desc *desc;
	struct vq_desc_extra *dxp;
	uint16_t i, prev_head, new_head;

	assert(vq->req_start != VQ_RING_DESC_CHAIN_END);
	assert(iovcnt <= vq->vq_free_cnt);

	/* TODO use indirect descriptors if iovcnt is high enough
	 * or the caller specifies SPDK_VIRTIO_DESC_F_INDIRECT
	 */

	prev_head = vq->req_end;
	new_head = vq->vq_desc_head_idx;
	for (i = 0; i < iovcnt; ++i) {
		desc = &vq->vq_ring.desc[new_head];

		if (!vq->vdev->is_hw) {
			desc->addr  = (uintptr_t)iovs[i].iov_base;
		} else {
			desc->addr = spdk_vtophys(iovs[i].iov_base);
		}

		desc->len = iovs[i].iov_len;
		/* always set NEXT flag. unset it on the last descriptor
		 * in the request-ending function.
		 */
		desc->flags = desc_type | VRING_DESC_F_NEXT;

		prev_head = new_head;
		new_head = desc->next;
	}

	dxp = &vq->vq_descx[vq->req_start];
	dxp->ndescs += iovcnt;

	vq->req_end = prev_head;
	vq->vq_desc_head_idx = new_head;
	vq->vq_free_cnt = (uint16_t)(vq->vq_free_cnt - iovcnt);
	if (vq->vq_desc_head_idx == VQ_RING_DESC_CHAIN_END) {
		assert(vq->vq_free_cnt == 0);
		vq->vq_desc_tail_idx = VQ_RING_DESC_CHAIN_END;
	}
}

#define DESC_PER_CACHELINE (RTE_CACHE_LINE_SIZE / sizeof(struct vring_desc))
uint16_t
virtio_recv_pkts(struct virtqueue *vq, void **io, uint32_t *len, uint16_t nb_pkts)
{
	uint16_t nb_used, num;

	nb_used = vq->vq_ring.used->idx - vq->vq_used_cons_idx;
	virtio_rmb();

	num = (uint16_t)(spdk_likely(nb_used <= nb_pkts) ? nb_used : nb_pkts);
	if (spdk_likely(num > DESC_PER_CACHELINE)) {
		num = num - ((vq->vq_used_cons_idx + num) % DESC_PER_CACHELINE);
	}

	return virtqueue_dequeue_burst_rx(vq, io, len, num);
}

int
virtio_dev_acquire_queue(struct virtio_dev *vdev, uint16_t index)
{
	struct virtqueue *vq = NULL;

	if (index >= vdev->max_queues) {
		SPDK_ERRLOG("requested vq index %"PRIu16" exceeds max queue count %"PRIu16".\n",
			    index, vdev->max_queues);
		return -1;
	}

	pthread_mutex_lock(&vdev->mutex);
	vq = vdev->vqs[index];
	if (vq == NULL || vq->owner_thread != NULL) {
		pthread_mutex_unlock(&vdev->mutex);
		return -1;
	}

	vq->owner_thread = spdk_get_thread();
	pthread_mutex_unlock(&vdev->mutex);
	return 0;
}

int32_t
virtio_dev_find_and_acquire_queue(struct virtio_dev *vdev, uint16_t start_index)
{
	struct virtqueue *vq = NULL;
	uint16_t i;

	pthread_mutex_lock(&vdev->mutex);
	for (i = start_index; i < vdev->max_queues; ++i) {
		vq = vdev->vqs[i];
		if (vq != NULL && vq->owner_thread == NULL) {
			break;
		}
	}

	if (vq == NULL || i == vdev->max_queues) {
		SPDK_ERRLOG("no more unused virtio queues with idx >= %"PRIu16".\n", start_index);
		pthread_mutex_unlock(&vdev->mutex);
		return -1;
	}

	vq->owner_thread = spdk_get_thread();
	pthread_mutex_unlock(&vdev->mutex);
	return i;
}

struct spdk_thread *
virtio_dev_queue_get_thread(struct virtio_dev *vdev, uint16_t index)
{
	struct spdk_thread *thread = NULL;

	if (index >= vdev->max_queues) {
		SPDK_ERRLOG("given vq index %"PRIu16" exceeds max queue count %"PRIu16"\n",
			    index, vdev->max_queues);
		abort(); /* This is not recoverable */
	}

	pthread_mutex_lock(&vdev->mutex);
	thread = vdev->vqs[index]->owner_thread;
	pthread_mutex_unlock(&vdev->mutex);

	return thread;
}

bool
virtio_dev_queue_is_acquired(struct virtio_dev *vdev, uint16_t index)
{
	return virtio_dev_queue_get_thread(vdev, index) != NULL;
}

void
virtio_dev_release_queue(struct virtio_dev *vdev, uint16_t index)
{
	struct virtqueue *vq = NULL;

	if (index >= vdev->max_queues) {
		SPDK_ERRLOG("given vq index %"PRIu16" exceeds max queue count %"PRIu16".\n",
			    index, vdev->max_queues);
		return;
	}

	pthread_mutex_lock(&vdev->mutex);
	vq = vdev->vqs[index];
	if (vq == NULL) {
		SPDK_ERRLOG("virtqueue at index %"PRIu16" is not initialized.\n", index);
		pthread_mutex_unlock(&vdev->mutex);
		return;
	}

	assert(vq->owner_thread == spdk_get_thread());
	vq->owner_thread = NULL;
	pthread_mutex_unlock(&vdev->mutex);
}

int
virtio_dev_read_dev_config(struct virtio_dev *dev, size_t offset,
			   void *dst, int length)
{
	return virtio_dev_backend_ops(dev)->read_dev_cfg(dev, offset, dst, length);
}

int
virtio_dev_write_dev_config(struct virtio_dev *dev, size_t offset,
			    const void *src, int length)
{
	return virtio_dev_backend_ops(dev)->write_dev_cfg(dev, offset, src, length);
}

void
virtio_dev_stop(struct virtio_dev *dev)
{
	virtio_dev_backend_ops(dev)->set_status(dev, VIRTIO_CONFIG_S_RESET);
	/* flush status write */
	virtio_dev_backend_ops(dev)->get_status(dev);
	virtio_free_queues(dev);
}

void
virtio_dev_set_status(struct virtio_dev *dev, uint8_t status)
{
	if (status != VIRTIO_CONFIG_S_RESET) {
		status |= virtio_dev_backend_ops(dev)->get_status(dev);
	}

	virtio_dev_backend_ops(dev)->set_status(dev, status);
}

uint8_t
virtio_dev_get_status(struct virtio_dev *dev)
{
	return virtio_dev_backend_ops(dev)->get_status(dev);
}

const struct virtio_dev_ops *
virtio_dev_backend_ops(struct virtio_dev *dev)
{
	return dev->backend_ops;
}

void
virtio_dev_dump_json_info(struct virtio_dev *hw, struct spdk_json_write_ctx *w)
{
	spdk_json_write_name(w, "virtio");
	spdk_json_write_object_begin(w);

	spdk_json_write_name(w, "vq_count");
	spdk_json_write_uint32(w, hw->max_queues);

	spdk_json_write_name(w, "vq_size");
	spdk_json_write_uint32(w, virtio_dev_backend_ops(hw)->get_queue_size(hw, 0));

	virtio_dev_backend_ops(hw)->dump_json_info(hw, w);

	spdk_json_write_object_end(w);
}

SPDK_LOG_REGISTER_COMPONENT("virtio_dev", SPDK_LOG_VIRTIO_DEV)