summaryrefslogtreecommitdiff
path: root/sys/dev/pci/virtio_pci.c
blob: 10e4ddd0b03d6098e4fc2e69e0170da0cc23aadb (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
/*	$OpenBSD: virtio_pci.c,v 1.26 2019/05/26 15:20:04 sf Exp $	*/
/*	$NetBSD: virtio.c,v 1.3 2011/11/02 23:05:52 njoly Exp $	*/

/*
 * Copyright (c) 2012 Stefan Fritsch.
 * Copyright (c) 2010 Minoura Makoto.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. 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.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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 <sys/param.h>
#include <sys/systm.h>
#include <sys/device.h>
#include <sys/mutex.h>

#include <dev/pci/pcidevs.h>
#include <dev/pci/pcireg.h>
#include <dev/pci/pcivar.h>

#include <dev/pv/virtioreg.h>
#include <dev/pv/virtiovar.h>
#include <dev/pci/virtio_pcireg.h>

/*
 * XXX: Before being used on big endian arches, the access to config registers
 * XXX: needs to be reviewed/fixed. The non-device specific registers are
 * XXX: PCI-endian while the device specific registers are native endian.
 */

#define MAX_MSIX_VECS	8

struct virtio_pci_softc;

int		virtio_pci_match(struct device *, void *, void *);
void		virtio_pci_attach(struct device *, struct device *, void *);
int		virtio_pci_detach(struct device *, int);

void		virtio_pci_kick(struct virtio_softc *, uint16_t);
int		virtio_pci_adjust_config_region(struct virtio_pci_softc *);
uint8_t		virtio_pci_read_device_config_1(struct virtio_softc *, int);
uint16_t	virtio_pci_read_device_config_2(struct virtio_softc *, int);
uint32_t	virtio_pci_read_device_config_4(struct virtio_softc *, int);
uint64_t	virtio_pci_read_device_config_8(struct virtio_softc *, int);
void		virtio_pci_write_device_config_1(struct virtio_softc *, int, uint8_t);
void		virtio_pci_write_device_config_2(struct virtio_softc *, int, uint16_t);
void		virtio_pci_write_device_config_4(struct virtio_softc *, int, uint32_t);
void		virtio_pci_write_device_config_8(struct virtio_softc *, int, uint64_t);
uint16_t	virtio_pci_read_queue_size(struct virtio_softc *, uint16_t);
void		virtio_pci_setup_queue(struct virtio_softc *, struct virtqueue *, uint64_t);
void		virtio_pci_set_status(struct virtio_softc *, int);
int		virtio_pci_negotiate_features(struct virtio_softc *,
    const struct virtio_feature_name *);
void		virtio_pci_set_msix_queue_vector(struct virtio_pci_softc *, uint32_t, uint16_t);
void		virtio_pci_set_msix_config_vector(struct virtio_pci_softc *, uint16_t);
int		virtio_pci_msix_establish(struct virtio_pci_softc *, struct pci_attach_args *, int, int (*)(void *), void *);
int		virtio_pci_setup_msix(struct virtio_pci_softc *, struct pci_attach_args *, int);
void		virtio_pci_free_irqs(struct virtio_pci_softc *);
int		virtio_pci_poll_intr(void *);
int		virtio_pci_legacy_intr(void *);
int		virtio_pci_legacy_intr_mpsafe(void *);
int		virtio_pci_config_intr(void *);
int		virtio_pci_queue_intr(void *);
int		virtio_pci_shared_queue_intr(void *);

enum irq_type {
	IRQ_NO_MSIX,
	IRQ_MSIX_SHARED, /* vec 0: config irq, vec 1 shared by all vqs */
	IRQ_MSIX_PER_VQ, /* vec 0: config irq, vec n: irq of vq[n-1] */
};

struct virtio_pci_softc {
	struct virtio_softc	sc_sc;
	pci_chipset_tag_t	sc_pc;
	pcitag_t		sc_ptag;

	bus_space_tag_t		sc_iot;
	bus_space_handle_t	sc_ioh;
	bus_size_t		sc_iosize;

	bus_space_tag_t		sc_notify_iot;
	bus_space_handle_t	sc_notify_ioh;
	bus_size_t		sc_notify_iosize;

	bus_space_tag_t		sc_devcfg_iot;
	bus_space_handle_t	sc_devcfg_ioh;
	bus_size_t		sc_devcfg_iosize;
	/*
	 * With 0.9, the offset of the devcfg region in the io bar changes
	 * depending on MSI-X being enabled or not.
	 * With 1.0, this field is still used to remember if MSI-X is enabled
	 * or not.
	 */
	unsigned int		sc_devcfg_offset;

	bus_space_tag_t		sc_isr_iot;
	bus_space_handle_t	sc_isr_ioh;
	bus_size_t		sc_isr_iosize;

	void			*sc_ih[MAX_MSIX_VECS];

	enum irq_type		sc_irq_type;
};

struct cfattach virtio_pci_ca = {
	sizeof(struct virtio_pci_softc),
	virtio_pci_match,
	virtio_pci_attach,
	virtio_pci_detach,
	NULL
};

struct virtio_ops virtio_pci_ops = {
	virtio_pci_kick,
	virtio_pci_read_device_config_1,
	virtio_pci_read_device_config_2,
	virtio_pci_read_device_config_4,
	virtio_pci_read_device_config_8,
	virtio_pci_write_device_config_1,
	virtio_pci_write_device_config_2,
	virtio_pci_write_device_config_4,
	virtio_pci_write_device_config_8,
	virtio_pci_read_queue_size,
	virtio_pci_setup_queue,
	virtio_pci_set_status,
	virtio_pci_negotiate_features,
	virtio_pci_poll_intr,
};

uint16_t
virtio_pci_read_queue_size(struct virtio_softc *vsc, uint16_t idx)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_2(sc->sc_iot, sc->sc_ioh, VIRTIO_CONFIG_QUEUE_SELECT,
	    idx);
	return bus_space_read_2(sc->sc_iot, sc->sc_ioh,
	    VIRTIO_CONFIG_QUEUE_SIZE);
}

void
virtio_pci_setup_queue(struct virtio_softc *vsc, struct virtqueue *vq,
    uint64_t addr)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_2(sc->sc_iot, sc->sc_ioh, VIRTIO_CONFIG_QUEUE_SELECT,
	    vq->vq_index);
	bus_space_write_4(sc->sc_iot, sc->sc_ioh, VIRTIO_CONFIG_QUEUE_ADDRESS,
	    addr / VIRTIO_PAGE_SIZE);

	/*
	 * This path is only executed if this function is called after
	 * the child's attach function has finished. In other cases,
	 * it's done in virtio_pci_setup_msix().
	 */
	if (sc->sc_irq_type != IRQ_NO_MSIX) {
		int vec = 1;
		if (sc->sc_irq_type == IRQ_MSIX_PER_VQ)
		       vec += vq->vq_index;
		bus_space_write_2(sc->sc_iot, sc->sc_ioh,
		    VIRTIO_MSI_QUEUE_VECTOR, vec);
	}
}

void
virtio_pci_set_status(struct virtio_softc *vsc, int status)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	int old = 0;

	if (status != 0)
		old = bus_space_read_1(sc->sc_iot, sc->sc_ioh,
				       VIRTIO_CONFIG_DEVICE_STATUS);
	bus_space_write_1(sc->sc_iot, sc->sc_ioh, VIRTIO_CONFIG_DEVICE_STATUS,
			  status|old);
}

int
virtio_pci_match(struct device *parent, void *match, void *aux)
{
	struct pci_attach_args *pa;

	pa = (struct pci_attach_args *)aux;
	if (PCI_VENDOR(pa->pa_id) == PCI_VENDOR_QUMRANET &&
	    PCI_PRODUCT(pa->pa_id) >= 0x1000 &&
	    PCI_PRODUCT(pa->pa_id) <= 0x103f &&
	    PCI_REVISION(pa->pa_class) == 0)
		return 1;
	if (PCI_VENDOR(pa->pa_id) == PCI_VENDOR_OPENBSD &&
	    PCI_PRODUCT(pa->pa_id) == PCI_PRODUCT_OPENBSD_CONTROL)
		return 1;
	return 0;
}

void
virtio_pci_attach(struct device *parent, struct device *self, void *aux)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)self;
	struct virtio_softc *vsc = &sc->sc_sc;
	struct pci_attach_args *pa = (struct pci_attach_args *)aux;
	pci_chipset_tag_t pc = pa->pa_pc;
	pcitag_t tag = pa->pa_tag;
	int revision;
	pcireg_t id;
	char const *intrstr;
	pci_intr_handle_t ih;

	revision = PCI_REVISION(pa->pa_class);
	if (revision != 0) {
		printf("unknown revision 0x%02x; giving up\n", revision);
		return;
	}

	/* subsystem ID shows what I am */
	id = PCI_PRODUCT(pci_conf_read(pc, tag, PCI_SUBSYS_ID_REG));

	printf("\n");

	vsc->sc_ops = &virtio_pci_ops;
	sc->sc_pc = pc;
	sc->sc_ptag = pa->pa_tag;
	vsc->sc_dmat = pa->pa_dmat;

	/*
	 * For virtio, ignore normal MSI black/white-listing depending on the
	 * PCI bridge but enable it unconditionally.
	 */
	pa->pa_flags |= PCI_FLAGS_MSI_ENABLED;

	if (pci_mapreg_map(pa, PCI_MAPREG_START, PCI_MAPREG_TYPE_IO, 0,
	    &sc->sc_iot, &sc->sc_ioh, NULL, &sc->sc_iosize, 0)) {
		printf("%s: can't map i/o space\n", vsc->sc_dev.dv_xname);
		return;
	}

	if (bus_space_subregion(sc->sc_iot, sc->sc_ioh,
	    VIRTIO_CONFIG_QUEUE_NOTIFY, 2, &sc->sc_notify_ioh) != 0) {
		printf("%s: can't map notify i/o space\n",
		    vsc->sc_dev.dv_xname);
		return;
	}
	sc->sc_notify_iosize = 2;
	sc->sc_notify_iot = sc->sc_iot;

	if (bus_space_subregion(sc->sc_iot, sc->sc_ioh,
	    VIRTIO_CONFIG_ISR_STATUS, 1, &sc->sc_isr_ioh) != 0) {
		printf("%s: can't map isr i/o space\n",
		    vsc->sc_dev.dv_xname);
		return;
	}
	sc->sc_isr_iosize = 1;
	sc->sc_isr_iot = sc->sc_iot;

	sc->sc_devcfg_offset = VIRTIO_CONFIG_DEVICE_CONFIG_NOMSI;
	sc->sc_irq_type = IRQ_NO_MSIX;
	if (virtio_pci_adjust_config_region(sc) != 0)
		return;

	virtio_device_reset(vsc);
	virtio_pci_set_status(vsc, VIRTIO_CONFIG_DEVICE_STATUS_ACK);
	virtio_pci_set_status(vsc, VIRTIO_CONFIG_DEVICE_STATUS_DRIVER);

	vsc->sc_childdevid = id;
	vsc->sc_child = NULL;
	config_found(self, sc, NULL);
	if (vsc->sc_child == NULL) {
		printf("%s: no matching child driver; not configured\n",
		    vsc->sc_dev.dv_xname);
		goto fail_1;
	}
	if (vsc->sc_child == VIRTIO_CHILD_ERROR) {
		printf("%s: virtio configuration failed\n",
		    vsc->sc_dev.dv_xname);
		goto fail_1;
	}

	if (virtio_pci_setup_msix(sc, pa, 0) == 0) {
		sc->sc_irq_type = IRQ_MSIX_PER_VQ;
		intrstr = "msix per-VQ";
	} else if (virtio_pci_setup_msix(sc, pa, 1) == 0) {
		sc->sc_irq_type = IRQ_MSIX_SHARED;
		intrstr = "msix shared";
	} else {
		int (*ih_func)(void *) = virtio_pci_legacy_intr;
		if (pci_intr_map_msi(pa, &ih) != 0 && pci_intr_map(pa, &ih) != 0) {
			printf("%s: couldn't map interrupt\n", vsc->sc_dev.dv_xname);
			goto fail_2;
		}
		intrstr = pci_intr_string(pc, ih);
		/*
		 * We always set the IPL_MPSAFE flag in order to do the relatively
		 * expensive ISR read without lock, and then grab the kernel lock in
		 * the interrupt handler.
		 */
		if (vsc->sc_ipl & IPL_MPSAFE)
			ih_func = virtio_pci_legacy_intr_mpsafe;
		sc->sc_ih[0] = pci_intr_establish(pc, ih, vsc->sc_ipl | IPL_MPSAFE,
		    ih_func, sc, vsc->sc_dev.dv_xname);
		if (sc->sc_ih[0] == NULL) {
			printf("%s: couldn't establish interrupt", vsc->sc_dev.dv_xname);
			if (intrstr != NULL)
				printf(" at %s", intrstr);
			printf("\n");
			goto fail_2;
		}
	}
	printf("%s: %s\n", vsc->sc_dev.dv_xname, intrstr);

	virtio_set_status(vsc, VIRTIO_CONFIG_DEVICE_STATUS_DRIVER_OK);
	return;

fail_2:
	config_detach(vsc->sc_child, 0);
fail_1:
	/* no pci_mapreg_unmap() or pci_intr_unmap() */
	virtio_set_status(vsc, VIRTIO_CONFIG_DEVICE_STATUS_FAILED);
}

int
virtio_pci_detach(struct device *self, int flags)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)self;
	struct virtio_softc *vsc = &sc->sc_sc;
	int r;

	if (vsc->sc_child != 0 && vsc->sc_child != VIRTIO_CHILD_ERROR) {
		r = config_detach(vsc->sc_child, flags);
		if (r)
			return r;
	}
	KASSERT(vsc->sc_child == 0 || vsc->sc_child == VIRTIO_CHILD_ERROR);
	KASSERT(vsc->sc_vqs == 0);
	virtio_pci_free_irqs(sc);
	if (sc->sc_iosize)
		bus_space_unmap(sc->sc_iot, sc->sc_ioh, sc->sc_iosize);
	sc->sc_iosize = 0;

	return 0;
}

int
virtio_pci_adjust_config_region(struct virtio_pci_softc *sc)
{
	sc->sc_devcfg_iosize = sc->sc_iosize - sc->sc_devcfg_offset;
	sc->sc_devcfg_iot = sc->sc_iot;
	if (bus_space_subregion(sc->sc_iot, sc->sc_ioh, sc->sc_devcfg_offset,
	    sc->sc_devcfg_iosize, &sc->sc_devcfg_ioh) != 0) {
		printf("%s: can't map config i/o space\n",
		    sc->sc_sc.sc_dev.dv_xname);
		return 1;
	}
	return 0;
}

/*
 * Feature negotiation.
 * Prints available / negotiated features if guest_feature_names != NULL and
 * VIRTIO_DEBUG is 1
 */
int
virtio_pci_negotiate_features(struct virtio_softc *vsc,
    const struct virtio_feature_name *guest_feature_names)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	uint64_t host, neg;

	vsc->sc_active_features = 0;

	/*
	 * We enable indirect descriptors by default. They can be switched
	 * off by setting bit 1 in the driver flags, see config(8)
	 */
	if (!(vsc->sc_dev.dv_cfdata->cf_flags & VIRTIO_CF_NO_INDIRECT) &&
	    !(vsc->sc_child->dv_cfdata->cf_flags & VIRTIO_CF_NO_INDIRECT)) {
		vsc->sc_driver_features |= VIRTIO_F_RING_INDIRECT_DESC;
	} else if (guest_feature_names != NULL) {
		printf(" RingIndirectDesc disabled by UKC");
	}

	/*
	 * The driver must add VIRTIO_F_RING_EVENT_IDX if it supports it.
	 * If it did, check if it is disabled by bit 2 in the driver flags.
	 */
	if ((vsc->sc_driver_features & VIRTIO_F_RING_EVENT_IDX) &&
	    ((vsc->sc_dev.dv_cfdata->cf_flags & VIRTIO_CF_NO_EVENT_IDX) ||
	    (vsc->sc_child->dv_cfdata->cf_flags & VIRTIO_CF_NO_EVENT_IDX))) {
		if (guest_feature_names != NULL)
			printf(" RingEventIdx disabled by UKC");
		vsc->sc_driver_features &= ~(VIRTIO_F_RING_EVENT_IDX);
	}

	host = bus_space_read_4(sc->sc_iot, sc->sc_ioh,
				VIRTIO_CONFIG_DEVICE_FEATURES);
	neg = host & vsc->sc_driver_features;
#if VIRTIO_DEBUG
	if (guest_feature_names)
		virtio_log_features(host, neg, guest_feature_names);
#endif
	bus_space_write_4(sc->sc_iot, sc->sc_ioh,
			  VIRTIO_CONFIG_GUEST_FEATURES, neg);
	vsc->sc_active_features = neg;
	if (neg & VIRTIO_F_RING_INDIRECT_DESC)
		vsc->sc_indirect = 1;
	else
		vsc->sc_indirect = 0;

	return 0;
}

/*
 * Device configuration registers.
 */
uint8_t
virtio_pci_read_device_config_1(struct virtio_softc *vsc, int index)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	return bus_space_read_1(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index);
}

uint16_t
virtio_pci_read_device_config_2(struct virtio_softc *vsc, int index)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	return bus_space_read_2(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index);
}

uint32_t
virtio_pci_read_device_config_4(struct virtio_softc *vsc, int index)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	return bus_space_read_4(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index);
}

uint64_t
virtio_pci_read_device_config_8(struct virtio_softc *vsc, int index)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	uint64_t r;

	r = bus_space_read_4(sc->sc_devcfg_iot, sc->sc_devcfg_ioh,
	    index + sizeof(uint32_t));
	r <<= 32;
	r += bus_space_read_4(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index);
	return r;
}

void
virtio_pci_write_device_config_1(struct virtio_softc *vsc, int index,
    uint8_t value)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_1(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index, value);
}

void
virtio_pci_write_device_config_2(struct virtio_softc *vsc, int index,
    uint16_t value)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_2(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index, value);
}

void
virtio_pci_write_device_config_4(struct virtio_softc *vsc,
			     int index, uint32_t value)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_4(sc->sc_devcfg_iot, sc->sc_devcfg_ioh, index, value);
}

void
virtio_pci_write_device_config_8(struct virtio_softc *vsc,
			     int index, uint64_t value)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_4(sc->sc_devcfg_iot, sc->sc_devcfg_ioh,
	    index, value & 0xffffffff);
	bus_space_write_4(sc->sc_devcfg_iot, sc->sc_devcfg_ioh,
	    index + sizeof(uint32_t), value >> 32);
}

int
virtio_pci_msix_establish(struct virtio_pci_softc *sc,
    struct pci_attach_args *pa, int idx, int (*handler)(void *), void *ih_arg)
{
	struct virtio_softc *vsc = &sc->sc_sc;
	pci_intr_handle_t ih;

	if (pci_intr_map_msix(pa, idx, &ih) != 0) {
#if VIRTIO_DEBUG
		printf("%s[%d]: pci_intr_map_msix failed\n",
		    vsc->sc_dev.dv_xname, idx);
#endif
		return 1;
	}
	sc->sc_ih[idx] = pci_intr_establish(sc->sc_pc, ih, vsc->sc_ipl,
	    handler, ih_arg, vsc->sc_dev.dv_xname);
	if (sc->sc_ih[idx] == NULL) {
		printf("%s[%d]: couldn't establish msix interrupt\n",
		    vsc->sc_dev.dv_xname, idx);
		return 1;
	}
	return 0;
}

void
virtio_pci_set_msix_queue_vector(struct virtio_pci_softc *sc, uint32_t idx, uint16_t vector)
{
	bus_space_write_2(sc->sc_iot, sc->sc_ioh,
	    VIRTIO_CONFIG_QUEUE_SELECT, idx);
	bus_space_write_2(sc->sc_iot, sc->sc_ioh,
	    VIRTIO_MSI_QUEUE_VECTOR, vector);
}

void
virtio_pci_set_msix_config_vector(struct virtio_pci_softc *sc, uint16_t vector)
{
	bus_space_write_2(sc->sc_iot, sc->sc_ioh,
	    VIRTIO_MSI_CONFIG_VECTOR, vector);
}

void
virtio_pci_free_irqs(struct virtio_pci_softc *sc)
{
	struct virtio_softc *vsc = &sc->sc_sc;
	int i;

	if (sc->sc_devcfg_offset == VIRTIO_CONFIG_DEVICE_CONFIG_MSI) {
		for (i = 0; i < vsc->sc_nvqs; i++) {
			virtio_pci_set_msix_queue_vector(sc, i,
			    VIRTIO_MSI_NO_VECTOR);
		}
	}

	for (i = 0; i < MAX_MSIX_VECS; i++) {
		if (sc->sc_ih[i]) {
			pci_intr_disestablish(sc->sc_pc, sc->sc_ih[i]);
			sc->sc_ih[i] = NULL;
		}
	}

	sc->sc_devcfg_offset = VIRTIO_CONFIG_DEVICE_CONFIG_NOMSI;
	virtio_pci_adjust_config_region(sc);
}

int
virtio_pci_setup_msix(struct virtio_pci_softc *sc, struct pci_attach_args *pa,
    int shared)
{
	struct virtio_softc *vsc = &sc->sc_sc;
	int i;

	if (virtio_pci_msix_establish(sc, pa, 0, virtio_pci_config_intr, vsc))
		return 1;
	sc->sc_devcfg_offset = VIRTIO_CONFIG_DEVICE_CONFIG_MSI;
	virtio_pci_adjust_config_region(sc);
	virtio_pci_set_msix_config_vector(sc, 0);

	if (shared) {
		if (virtio_pci_msix_establish(sc, pa, 1,
		    virtio_pci_shared_queue_intr, vsc)) {
			goto fail;
		}

		for (i = 0; i < vsc->sc_nvqs; i++)
			virtio_pci_set_msix_queue_vector(sc, i, 1);
	} else {
		for (i = 0; i <= vsc->sc_nvqs; i++) {
			if (virtio_pci_msix_establish(sc, pa, i + 1,
			    virtio_pci_queue_intr, &vsc->sc_vqs[i])) {
				goto fail;
			}
			virtio_pci_set_msix_queue_vector(sc, i, i + 1);
		}
	}

	return 0;
fail:
	virtio_pci_free_irqs(sc);
	return 1;
}

/*
 * Interrupt handler.
 */

/*
 * Only used without MSI-X
 */
int
virtio_pci_legacy_intr(void *arg)
{
	struct virtio_pci_softc *sc = arg;
	struct virtio_softc *vsc = &sc->sc_sc;
	int isr, r = 0;

	/* check and ack the interrupt */
	isr = bus_space_read_1(sc->sc_isr_iot, sc->sc_isr_ioh, 0);
	if (isr == 0)
		return 0;
	KERNEL_LOCK();
	if ((isr & VIRTIO_CONFIG_ISR_CONFIG_CHANGE) &&
	    (vsc->sc_config_change != NULL)) {
		r = (vsc->sc_config_change)(vsc);
	}
	r |= virtio_check_vqs(vsc);
	KERNEL_UNLOCK();

	return r;
}

int
virtio_pci_legacy_intr_mpsafe(void *arg)
{
	struct virtio_pci_softc *sc = arg;
	struct virtio_softc *vsc = &sc->sc_sc;
	int isr, r = 0;

	/* check and ack the interrupt */
	isr = bus_space_read_1(sc->sc_isr_iot, sc->sc_isr_ioh, 0);
	if (isr == 0)
		return 0;
	if ((isr & VIRTIO_CONFIG_ISR_CONFIG_CHANGE) &&
	    (vsc->sc_config_change != NULL)) {
		r = (vsc->sc_config_change)(vsc);
	}
	r |= virtio_check_vqs(vsc);
	return r;
}

/*
 * Only used with MSI-X
 */
int
virtio_pci_config_intr(void *arg)
{
	struct virtio_softc *vsc = arg;

	if (vsc->sc_config_change != NULL)
		return vsc->sc_config_change(vsc);
	return 0;
}

/*
 * Only used with MSI-X
 */
int
virtio_pci_queue_intr(void *arg)
{
	struct virtqueue *vq = arg;

	if (vq->vq_done)
		return (vq->vq_done)(vq);
	return 0;
}

int
virtio_pci_shared_queue_intr(void *arg)
{
	struct virtio_softc *vsc = arg;

	return virtio_check_vqs(vsc);
}

/*
 * Interrupt handler to be used when polling.
 * We cannot use isr here because it is not defined in MSI-X mode.
 */
int
virtio_pci_poll_intr(void *arg)
{
	struct virtio_pci_softc *sc = arg;
	struct virtio_softc *vsc = &sc->sc_sc;
	int r = 0;

	if (vsc->sc_config_change != NULL)
		r = (vsc->sc_config_change)(vsc);

	r |= virtio_check_vqs(vsc);

	return r;
}

void
virtio_pci_kick(struct virtio_softc *vsc, uint16_t idx)
{
	struct virtio_pci_softc *sc = (struct virtio_pci_softc *)vsc;
	bus_space_write_2(sc->sc_notify_iot, sc->sc_notify_ioh, 0, idx);
}