summaryrefslogtreecommitdiffstats
path: root/arch/um/drivers/virtio_uml.c
blob: 77faa2cf3a13f0c1c6917e412db66f25a64311d7 (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
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
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * Virtio vhost-user driver
 *
 * Copyright(c) 2019 Intel Corporation
 *
 * This driver allows virtio devices to be used over a vhost-user socket.
 *
 * Guest devices can be instantiated by kernel module or command line
 * parameters. One device will be created for each parameter. Syntax:
 *
 *		virtio_uml.device=<socket>:<virtio_id>[:<platform_id>]
 * where:
 *		<socket>	:= vhost-user socket path to connect
 *		<virtio_id>	:= virtio device id (as in virtio_ids.h)
 *		<platform_id>	:= (optional) platform device id
 *
 * example:
 *		virtio_uml.device=/var/uml.socket:1
 *
 * Based on Virtio MMIO driver by Pawel Moll, copyright 2011-2014, ARM Ltd.
 */
#include <linux/module.h>
#include <linux/of.h>
#include <linux/platform_device.h>
#include <linux/slab.h>
#include <linux/virtio.h>
#include <linux/virtio_config.h>
#include <linux/virtio_ring.h>
#include <linux/time-internal.h>
#include <linux/virtio-uml.h>
#include <shared/as-layout.h>
#include <irq_kern.h>
#include <init.h>
#include <os.h>
#include "vhost_user.h"

#define MAX_SUPPORTED_QUEUE_SIZE	256

#define to_virtio_uml_device(_vdev) \
	container_of(_vdev, struct virtio_uml_device, vdev)

struct virtio_uml_platform_data {
	u32 virtio_device_id;
	const char *socket_path;
	struct work_struct conn_broken_wk;
	struct platform_device *pdev;
};

struct virtio_uml_device {
	struct virtio_device vdev;
	struct platform_device *pdev;
	struct virtio_uml_platform_data *pdata;

	spinlock_t sock_lock;
	int sock, req_fd, irq;
	u64 features;
	u64 protocol_features;
	u8 status;
	u8 registered:1;
	u8 suspended:1;
	u8 no_vq_suspend:1;

	u8 config_changed_irq:1;
	uint64_t vq_irq_vq_map;
	int recv_rc;
};

struct virtio_uml_vq_info {
	int kick_fd, call_fd;
	char name[32];
	bool suspended;
};

extern unsigned long long physmem_size, highmem;

#define vu_err(vu_dev, ...)	dev_err(&(vu_dev)->pdev->dev, ##__VA_ARGS__)

/* Vhost-user protocol */

static int full_sendmsg_fds(int fd, const void *buf, unsigned int len,
			    const int *fds, unsigned int fds_num)
{
	int rc;

	do {
		rc = os_sendmsg_fds(fd, buf, len, fds, fds_num);
		if (rc > 0) {
			buf += rc;
			len -= rc;
			fds = NULL;
			fds_num = 0;
		}
	} while (len && (rc >= 0 || rc == -EINTR));

	if (rc < 0)
		return rc;
	return 0;
}

static int full_read(int fd, void *buf, int len, bool abortable)
{
	int rc;

	if (!len)
		return 0;

	do {
		rc = os_read_file(fd, buf, len);
		if (rc > 0) {
			buf += rc;
			len -= rc;
		}
	} while (len && (rc > 0 || rc == -EINTR || (!abortable && rc == -EAGAIN)));

	if (rc < 0)
		return rc;
	if (rc == 0)
		return -ECONNRESET;
	return 0;
}

static int vhost_user_recv_header(int fd, struct vhost_user_msg *msg)
{
	return full_read(fd, msg, sizeof(msg->header), true);
}

static int vhost_user_recv(struct virtio_uml_device *vu_dev,
			   int fd, struct vhost_user_msg *msg,
			   size_t max_payload_size, bool wait)
{
	size_t size;
	int rc;

	/*
	 * In virtio time-travel mode, we're handling all the vhost-user
	 * FDs by polling them whenever appropriate. However, we may get
	 * into a situation where we're sending out an interrupt message
	 * to a device (e.g. a net device) and need to handle a simulation
	 * time message while doing so, e.g. one that tells us to update
	 * our idea of how long we can run without scheduling.
	 *
	 * Thus, we need to not just read() from the given fd, but need
	 * to also handle messages for the simulation time - this function
	 * does that for us while waiting for the given fd to be readable.
	 */
	if (wait)
		time_travel_wait_readable(fd);

	rc = vhost_user_recv_header(fd, msg);

	if (rc)
		return rc;
	size = msg->header.size;
	if (size > max_payload_size)
		return -EPROTO;
	return full_read(fd, &msg->payload, size, false);
}

static void vhost_user_check_reset(struct virtio_uml_device *vu_dev,
				   int rc)
{
	struct virtio_uml_platform_data *pdata = vu_dev->pdata;

	if (rc != -ECONNRESET)
		return;

	if (!vu_dev->registered)
		return;

	vu_dev->registered = 0;

	schedule_work(&pdata->conn_broken_wk);
}

static int vhost_user_recv_resp(struct virtio_uml_device *vu_dev,
				struct vhost_user_msg *msg,
				size_t max_payload_size)
{
	int rc = vhost_user_recv(vu_dev, vu_dev->sock, msg,
				 max_payload_size, true);

	if (rc) {
		vhost_user_check_reset(vu_dev, rc);
		return rc;
	}

	if (msg->header.flags != (VHOST_USER_FLAG_REPLY | VHOST_USER_VERSION))
		return -EPROTO;

	return 0;
}

static int vhost_user_recv_u64(struct virtio_uml_device *vu_dev,
			       u64 *value)
{
	struct vhost_user_msg msg;
	int rc = vhost_user_recv_resp(vu_dev, &msg,
				      sizeof(msg.payload.integer));

	if (rc)
		return rc;
	if (msg.header.size != sizeof(msg.payload.integer))
		return -EPROTO;
	*value = msg.payload.integer;
	return 0;
}

static int vhost_user_recv_req(struct virtio_uml_device *vu_dev,
			       struct vhost_user_msg *msg,
			       size_t max_payload_size)
{
	int rc = vhost_user_recv(vu_dev, vu_dev->req_fd, msg,
				 max_payload_size, false);

	if (rc)
		return rc;

	if ((msg->header.flags & ~VHOST_USER_FLAG_NEED_REPLY) !=
			VHOST_USER_VERSION)
		return -EPROTO;

	return 0;
}

static int vhost_user_send(struct virtio_uml_device *vu_dev,
			   bool need_response, struct vhost_user_msg *msg,
			   int *fds, size_t num_fds)
{
	size_t size = sizeof(msg->header) + msg->header.size;
	unsigned long flags;
	bool request_ack;
	int rc;

	msg->header.flags |= VHOST_USER_VERSION;

	/*
	 * The need_response flag indicates that we already need a response,
	 * e.g. to read the features. In these cases, don't request an ACK as
	 * it is meaningless. Also request an ACK only if supported.
	 */
	request_ack = !need_response;
	if (!(vu_dev->protocol_features &
			BIT_ULL(VHOST_USER_PROTOCOL_F_REPLY_ACK)))
		request_ack = false;

	if (request_ack)
		msg->header.flags |= VHOST_USER_FLAG_NEED_REPLY;

	spin_lock_irqsave(&vu_dev->sock_lock, flags);
	rc = full_sendmsg_fds(vu_dev->sock, msg, size, fds, num_fds);
	if (rc < 0)
		goto out;

	if (request_ack) {
		uint64_t status;

		rc = vhost_user_recv_u64(vu_dev, &status);
		if (rc)
			goto out;

		if (status) {
			vu_err(vu_dev, "slave reports error: %llu\n", status);
			rc = -EIO;
			goto out;
		}
	}

out:
	spin_unlock_irqrestore(&vu_dev->sock_lock, flags);
	return rc;
}

static int vhost_user_send_no_payload(struct virtio_uml_device *vu_dev,
				      bool need_response, u32 request)
{
	struct vhost_user_msg msg = {
		.header.request = request,
	};

	return vhost_user_send(vu_dev, need_response, &msg, NULL, 0);
}

static int vhost_user_send_no_payload_fd(struct virtio_uml_device *vu_dev,
					 u32 request, int fd)
{
	struct vhost_user_msg msg = {
		.header.request = request,
	};

	return vhost_user_send(vu_dev, false, &msg, &fd, 1);
}

static int vhost_user_send_u64(struct virtio_uml_device *vu_dev,
			       u32 request, u64 value)
{
	struct vhost_user_msg msg = {
		.header.request = request,
		.header.size = sizeof(msg.payload.integer),
		.payload.integer = value,
	};

	return vhost_user_send(vu_dev, false, &msg, NULL, 0);
}

static int vhost_user_set_owner(struct virtio_uml_device *vu_dev)
{
	return vhost_user_send_no_payload(vu_dev, false, VHOST_USER_SET_OWNER);
}

static int vhost_user_get_features(struct virtio_uml_device *vu_dev,
				   u64 *features)
{
	int rc = vhost_user_send_no_payload(vu_dev, true,
					    VHOST_USER_GET_FEATURES);

	if (rc)
		return rc;
	return vhost_user_recv_u64(vu_dev, features);
}

static int vhost_user_set_features(struct virtio_uml_device *vu_dev,
				   u64 features)
{
	return vhost_user_send_u64(vu_dev, VHOST_USER_SET_FEATURES, features);
}

static int vhost_user_get_protocol_features(struct virtio_uml_device *vu_dev,
					    u64 *protocol_features)
{
	int rc = vhost_user_send_no_payload(vu_dev, true,
			VHOST_USER_GET_PROTOCOL_FEATURES);

	if (rc)
		return rc;
	return vhost_user_recv_u64(vu_dev, protocol_features);
}

static int vhost_user_set_protocol_features(struct virtio_uml_device *vu_dev,
					    u64 protocol_features)
{
	return vhost_user_send_u64(vu_dev, VHOST_USER_SET_PROTOCOL_FEATURES,
				   protocol_features);
}

static void vhost_user_reply(struct virtio_uml_device *vu_dev,
			     struct vhost_user_msg *msg, int response)
{
	struct vhost_user_msg reply = {
		.payload.integer = response,
	};
	size_t size = sizeof(reply.header) + sizeof(reply.payload.integer);
	int rc;

	reply.header = msg->header;
	reply.header.flags &= ~VHOST_USER_FLAG_NEED_REPLY;
	reply.header.flags |= VHOST_USER_FLAG_REPLY;
	reply.header.size = sizeof(reply.payload.integer);

	rc = full_sendmsg_fds(vu_dev->req_fd, &reply, size, NULL, 0);

	if (rc)
		vu_err(vu_dev,
		       "sending reply to slave request failed: %d (size %zu)\n",
		       rc, size);
}

static irqreturn_t vu_req_read_message(struct virtio_uml_device *vu_dev,
				       struct time_travel_event *ev)
{
	struct virtqueue *vq;
	int response = 1;
	struct {
		struct vhost_user_msg msg;
		u8 extra_payload[512];
	} msg;
	int rc;
	irqreturn_t irq_rc = IRQ_NONE;

	while (1) {
		rc = vhost_user_recv_req(vu_dev, &msg.msg,
					 sizeof(msg.msg.payload) +
					 sizeof(msg.extra_payload));
		if (rc)
			break;

		switch (msg.msg.header.request) {
		case VHOST_USER_SLAVE_CONFIG_CHANGE_MSG:
			vu_dev->config_changed_irq = true;
			response = 0;
			break;
		case VHOST_USER_SLAVE_VRING_CALL:
			virtio_device_for_each_vq((&vu_dev->vdev), vq) {
				if (vq->index == msg.msg.payload.vring_state.index) {
					response = 0;
					vu_dev->vq_irq_vq_map |= BIT_ULL(vq->index);
					break;
				}
			}
			break;
		case VHOST_USER_SLAVE_IOTLB_MSG:
			/* not supported - VIRTIO_F_ACCESS_PLATFORM */
		case VHOST_USER_SLAVE_VRING_HOST_NOTIFIER_MSG:
			/* not supported - VHOST_USER_PROTOCOL_F_HOST_NOTIFIER */
		default:
			vu_err(vu_dev, "unexpected slave request %d\n",
			       msg.msg.header.request);
		}

		if (ev && !vu_dev->suspended)
			time_travel_add_irq_event(ev);

		if (msg.msg.header.flags & VHOST_USER_FLAG_NEED_REPLY)
			vhost_user_reply(vu_dev, &msg.msg, response);
		irq_rc = IRQ_HANDLED;
	}
	/* mask EAGAIN as we try non-blocking read until socket is empty */
	vu_dev->recv_rc = (rc == -EAGAIN) ? 0 : rc;
	return irq_rc;
}

static irqreturn_t vu_req_interrupt(int irq, void *data)
{
	struct virtio_uml_device *vu_dev = data;
	irqreturn_t ret = IRQ_HANDLED;

	if (!um_irq_timetravel_handler_used())
		ret = vu_req_read_message(vu_dev, NULL);

	if (vu_dev->recv_rc) {
		vhost_user_check_reset(vu_dev, vu_dev->recv_rc);
	} else if (vu_dev->vq_irq_vq_map) {
		struct virtqueue *vq;

		virtio_device_for_each_vq((&vu_dev->vdev), vq) {
			if (vu_dev->vq_irq_vq_map & BIT_ULL(vq->index))
				vring_interrupt(0 /* ignored */, vq);
		}
		vu_dev->vq_irq_vq_map = 0;
	} else if (vu_dev->config_changed_irq) {
		virtio_config_changed(&vu_dev->vdev);
		vu_dev->config_changed_irq = false;
	}

	return ret;
}

static void vu_req_interrupt_comm_handler(int irq, int fd, void *data,
					  struct time_travel_event *ev)
{
	vu_req_read_message(data, ev);
}

static int vhost_user_init_slave_req(struct virtio_uml_device *vu_dev)
{
	int rc, req_fds[2];

	/* Use a pipe for slave req fd, SIGIO is not supported for eventfd */
	rc = os_pipe(req_fds, true, true);
	if (rc < 0)
		return rc;
	vu_dev->req_fd = req_fds[0];

	rc = um_request_irq_tt(UM_IRQ_ALLOC, vu_dev->req_fd, IRQ_READ,
			       vu_req_interrupt, IRQF_SHARED,
			       vu_dev->pdev->name, vu_dev,
			       vu_req_interrupt_comm_handler);
	if (rc < 0)
		goto err_close;

	vu_dev->irq = rc;

	rc = vhost_user_send_no_payload_fd(vu_dev, VHOST_USER_SET_SLAVE_REQ_FD,
					   req_fds[1]);
	if (rc)
		goto err_free_irq;

	goto out;

err_free_irq:
	um_free_irq(vu_dev->irq, vu_dev);
err_close:
	os_close_file(req_fds[0]);
out:
	/* Close unused write end of request fds */
	os_close_file(req_fds[1]);
	return rc;
}

static int vhost_user_init(struct virtio_uml_device *vu_dev)
{
	int rc = vhost_user_set_owner(vu_dev);

	if (rc)
		return rc;
	rc = vhost_user_get_features(vu_dev, &vu_dev->features);
	if (rc)
		return rc;

	if (vu_dev->features & BIT_ULL(VHOST_USER_F_PROTOCOL_FEATURES)) {
		rc = vhost_user_get_protocol_features(vu_dev,
				&vu_dev->protocol_features);
		if (rc)
			return rc;
		vu_dev->protocol_features &= VHOST_USER_SUPPORTED_PROTOCOL_F;
		rc = vhost_user_set_protocol_features(vu_dev,
				vu_dev->protocol_features);
		if (rc)
			return rc;
	}

	if (vu_dev->protocol_features &
			BIT_ULL(VHOST_USER_PROTOCOL_F_SLAVE_REQ)) {
		rc = vhost_user_init_slave_req(vu_dev);
		if (rc)
			return rc;
	}

	return 0;
}

static void vhost_user_get_config(struct virtio_uml_device *vu_dev,
				  u32 offset, void *buf, u32 len)
{
	u32 cfg_size = offset + len;
	struct vhost_user_msg *msg;
	size_t payload_size = sizeof(msg->payload.config) + cfg_size;
	size_t msg_size = sizeof(msg->header) + payload_size;
	int rc;

	if (!(vu_dev->protocol_features &
	      BIT_ULL(VHOST_USER_PROTOCOL_F_CONFIG)))
		return;

	msg = kzalloc(msg_size, GFP_KERNEL);
	if (!msg)
		return;
	msg->header.request = VHOST_USER_GET_CONFIG;
	msg->header.size = payload_size;
	msg->payload.config.offset = 0;
	msg->payload.config.size = cfg_size;

	rc = vhost_user_send(vu_dev, true, msg, NULL, 0);
	if (rc) {
		vu_err(vu_dev, "sending VHOST_USER_GET_CONFIG failed: %d\n",
		       rc);
		goto free;
	}

	rc = vhost_user_recv_resp(vu_dev, msg, msg_size);
	if (rc) {
		vu_err(vu_dev,
		       "receiving VHOST_USER_GET_CONFIG response failed: %d\n",
		       rc);
		goto free;
	}

	if (msg->header.size != payload_size ||
	    msg->payload.config.size != cfg_size) {
		rc = -EPROTO;
		vu_err(vu_dev,
		       "Invalid VHOST_USER_GET_CONFIG sizes (payload %d expected %zu, config %u expected %u)\n",
		       msg->header.size, payload_size,
		       msg->payload.config.size, cfg_size);
		goto free;
	}
	memcpy(buf, msg->payload.config.payload + offset, len);

free:
	kfree(msg);
}

static void vhost_user_set_config(struct virtio_uml_device *vu_dev,
				  u32 offset, const void *buf, u32 len)
{
	struct vhost_user_msg *msg;
	size_t payload_size = sizeof(msg->payload.config) + len;
	size_t msg_size = sizeof(msg->header) + payload_size;
	int rc;

	if (!(vu_dev->protocol_features &
	      BIT_ULL(VHOST_USER_PROTOCOL_F_CONFIG)))
		return;

	msg = kzalloc(msg_size, GFP_KERNEL);
	if (!msg)
		return;
	msg->header.request = VHOST_USER_SET_CONFIG;
	msg->header.size = payload_size;
	msg->payload.config.offset = offset;
	msg->payload.config.size = len;
	memcpy(msg->payload.config.payload, buf, len);

	rc = vhost_user_send(vu_dev, false, msg, NULL, 0);
	if (rc)
		vu_err(vu_dev, "sending VHOST_USER_SET_CONFIG failed: %d\n",
		       rc);

	kfree(msg);
}

static int vhost_user_init_mem_region(u64 addr, u64 size, int *fd_out,
				      struct vhost_user_mem_region *region_out)
{
	unsigned long long mem_offset;
	int rc = phys_mapping(addr, &mem_offset);

	if (WARN(rc < 0, "phys_mapping of 0x%llx returned %d\n", addr, rc))
		return -EFAULT;
	*fd_out = rc;
	region_out->guest_addr = addr;
	region_out->user_addr = addr;
	region_out->size = size;
	region_out->mmap_offset = mem_offset;

	/* Ensure mapping is valid for the entire region */
	rc = phys_mapping(addr + size - 1, &mem_offset);
	if (WARN(rc != *fd_out, "phys_mapping of 0x%llx failed: %d != %d\n",
		 addr + size - 1, rc, *fd_out))
		return -EFAULT;
	return 0;
}

static int vhost_user_set_mem_table(struct virtio_uml_device *vu_dev)
{
	struct vhost_user_msg msg = {
		.header.request = VHOST_USER_SET_MEM_TABLE,
		.header.size = sizeof(msg.payload.mem_regions),
		.payload.mem_regions.num = 1,
	};
	unsigned long reserved = uml_reserved - uml_physmem;
	int fds[2];
	int rc;

	/*
	 * This is a bit tricky, see also the comment with setup_physmem().
	 *
	 * Essentially, setup_physmem() uses a file to mmap() our physmem,
	 * but the code and data we *already* have is omitted. To us, this
	 * is no difference, since they both become part of our address
	 * space and memory consumption. To somebody looking in from the
	 * outside, however, it is different because the part of our memory
	 * consumption that's already part of the binary (code/data) is not
	 * mapped from the file, so it's not visible to another mmap from
	 * the file descriptor.
	 *
	 * Thus, don't advertise this space to the vhost-user slave. This
	 * means that the slave will likely abort or similar when we give
	 * it an address from the hidden range, since it's not marked as
	 * a valid address, but at least that way we detect the issue and
	 * don't just have the slave read an all-zeroes buffer from the
	 * shared memory file, or write something there that we can never
	 * see (depending on the direction of the virtqueue traffic.)
	 *
	 * Since we usually don't want to use .text for virtio buffers,
	 * this effectively means that you cannot use
	 *  1) global variables, which are in the .bss and not in the shm
	 *     file-backed memory
	 *  2) the stack in some processes, depending on where they have
	 *     their stack (or maybe only no interrupt stack?)
	 *
	 * The stack is already not typically valid for DMA, so this isn't
	 * much of a restriction, but global variables might be encountered.
	 *
	 * It might be possible to fix it by copying around the data that's
	 * between bss_start and where we map the file now, but it's not
	 * something that you typically encounter with virtio drivers, so
	 * it didn't seem worthwhile.
	 */
	rc = vhost_user_init_mem_region(reserved, physmem_size - reserved,
					&fds[0],
					&msg.payload.mem_regions.regions[0]);

	if (rc < 0)
		return rc;
	if (highmem) {
		msg.payload.mem_regions.num++;
		rc = vhost_user_init_mem_region(__pa(end_iomem), highmem,
				&fds[1], &msg.payload.mem_regions.regions[1]);
		if (rc < 0)
			return rc;
	}

	return vhost_user_send(vu_dev, false, &msg, fds,
			       msg.payload.mem_regions.num);
}

static int vhost_user_set_vring_state(struct virtio_uml_device *vu_dev,
				      u32 request, u32 index, u32 num)
{
	struct vhost_user_msg msg = {
		.header.request = request,
		.header.size = sizeof(msg.payload.vring_state),
		.payload.vring_state.index = index,
		.payload.vring_state.num = num,
	};

	return vhost_user_send(vu_dev, false, &msg, NULL, 0);
}

static int vhost_user_set_vring_num(struct virtio_uml_device *vu_dev,
				    u32 index, u32 num)
{
	return vhost_user_set_vring_state(vu_dev, VHOST_USER_SET_VRING_NUM,
					  index, num);
}

static int vhost_user_set_vring_base(struct virtio_uml_device *vu_dev,
				     u32 index, u32 offset)
{
	return vhost_user_set_vring_state(vu_dev, VHOST_USER_SET_VRING_BASE,
					  index, offset);
}

static int vhost_user_set_vring_addr(struct virtio_uml_device *vu_dev,
				     u32 index, u64 desc, u64 used, u64 avail,
				     u64 log)
{
	struct vhost_user_msg msg = {
		.header.request = VHOST_USER_SET_VRING_ADDR,
		.header.size = sizeof(msg.payload.vring_addr),
		.payload.vring_addr.index = index,
		.payload.vring_addr.desc = desc,
		.payload.vring_addr.used = used,
		.payload.vring_addr.avail = avail,
		.payload.vring_addr.log = log,
	};

	return vhost_user_send(vu_dev, false, &msg, NULL, 0);
}

static int vhost_user_set_vring_fd(struct virtio_uml_device *vu_dev,
				   u32 request, int index, int fd)
{
	struct vhost_user_msg msg = {
		.header.request = request,
		.header.size = sizeof(msg.payload.integer),
		.payload.integer = index,
	};

	if (index & ~VHOST_USER_VRING_INDEX_MASK)
		return -EINVAL;
	if (fd < 0) {
		msg.payload.integer |= VHOST_USER_VRING_POLL_MASK;
		return vhost_user_send(vu_dev, false, &msg, NULL, 0);
	}
	return vhost_user_send(vu_dev, false, &msg, &fd, 1);
}

static int vhost_user_set_vring_call(struct virtio_uml_device *vu_dev,
				     int index, int fd)
{
	return vhost_user_set_vring_fd(vu_dev, VHOST_USER_SET_VRING_CALL,
				       index, fd);
}

static int vhost_user_set_vring_kick(struct virtio_uml_device *vu_dev,
				     int index, int fd)
{
	return vhost_user_set_vring_fd(vu_dev, VHOST_USER_SET_VRING_KICK,
				       index, fd);
}

static int vhost_user_set_vring_enable(struct virtio_uml_device *vu_dev,
				       u32 index, bool enable)
{
	if (!(vu_dev->features & BIT_ULL(VHOST_USER_F_PROTOCOL_FEATURES)))
		return 0;

	return vhost_user_set_vring_state(vu_dev, VHOST_USER_SET_VRING_ENABLE,
					  index, enable);
}


/* Virtio interface */

static bool vu_notify(struct virtqueue *vq)
{
	struct virtio_uml_vq_info *info = vq->priv;
	const uint64_t n = 1;
	int rc;

	if (info->suspended)
		return true;

	time_travel_propagate_time();

	if (info->kick_fd < 0) {
		struct virtio_uml_device *vu_dev;

		vu_dev = to_virtio_uml_device(vq->vdev);

		return vhost_user_set_vring_state(vu_dev, VHOST_USER_VRING_KICK,
						  vq->index, 0) == 0;
	}

	do {
		rc = os_write_file(info->kick_fd, &n, sizeof(n));
	} while (rc == -EINTR);
	return !WARN(rc != sizeof(n), "write returned %d\n", rc);
}

static irqreturn_t vu_interrupt(int irq, void *opaque)
{
	struct virtqueue *vq = opaque;
	struct virtio_uml_vq_info *info = vq->priv;
	uint64_t n;
	int rc;
	irqreturn_t ret = IRQ_NONE;

	do {
		rc = os_read_file(info->call_fd, &n, sizeof(n));
		if (rc == sizeof(n))
			ret |= vring_interrupt(irq, vq);
	} while (rc == sizeof(n) || rc == -EINTR);
	WARN(rc != -EAGAIN, "read returned %d\n", rc);
	return ret;
}


static void vu_get(struct virtio_device *vdev, unsigned offset,
		   void *buf, unsigned len)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	vhost_user_get_config(vu_dev, offset, buf, len);
}

static void vu_set(struct virtio_device *vdev, unsigned offset,
		   const void *buf, unsigned len)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	vhost_user_set_config(vu_dev, offset, buf, len);
}

static u8 vu_get_status(struct virtio_device *vdev)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	return vu_dev->status;
}

static void vu_set_status(struct virtio_device *vdev, u8 status)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	vu_dev->status = status;
}

static void vu_reset(struct virtio_device *vdev)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	vu_dev->status = 0;
}

static void vu_del_vq(struct virtqueue *vq)
{
	struct virtio_uml_vq_info *info = vq->priv;

	if (info->call_fd >= 0) {
		struct virtio_uml_device *vu_dev;

		vu_dev = to_virtio_uml_device(vq->vdev);

		um_free_irq(vu_dev->irq, vq);
		os_close_file(info->call_fd);
	}

	if (info->kick_fd >= 0)
		os_close_file(info->kick_fd);

	vring_del_virtqueue(vq);
	kfree(info);
}

static void vu_del_vqs(struct virtio_device *vdev)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);
	struct virtqueue *vq, *n;
	u64 features;

	/* Note: reverse order as a workaround to a decoding bug in snabb */
	list_for_each_entry_reverse(vq, &vdev->vqs, list)
		WARN_ON(vhost_user_set_vring_enable(vu_dev, vq->index, false));

	/* Ensure previous messages have been processed */
	WARN_ON(vhost_user_get_features(vu_dev, &features));

	list_for_each_entry_safe(vq, n, &vdev->vqs, list)
		vu_del_vq(vq);
}

static int vu_setup_vq_call_fd(struct virtio_uml_device *vu_dev,
			       struct virtqueue *vq)
{
	struct virtio_uml_vq_info *info = vq->priv;
	int call_fds[2];
	int rc;

	/* no call FD needed/desired in this case */
	if (vu_dev->protocol_features &
			BIT_ULL(VHOST_USER_PROTOCOL_F_INBAND_NOTIFICATIONS) &&
	    vu_dev->protocol_features &
			BIT_ULL(VHOST_USER_PROTOCOL_F_SLAVE_REQ)) {
		info->call_fd = -1;
		return 0;
	}

	/* Use a pipe for call fd, since SIGIO is not supported for eventfd */
	rc = os_pipe(call_fds, true, true);
	if (rc < 0)
		return rc;

	info->call_fd = call_fds[0];
	rc = um_request_irq(vu_dev->irq, info->call_fd, IRQ_READ,
			    vu_interrupt, IRQF_SHARED, info->name, vq);
	if (rc < 0)
		goto close_both;

	rc = vhost_user_set_vring_call(vu_dev, vq->index, call_fds[1]);
	if (rc)
		goto release_irq;

	goto out;

release_irq:
	um_free_irq(vu_dev->irq, vq);
close_both:
	os_close_file(call_fds[0]);
out:
	/* Close (unused) write end of call fds */
	os_close_file(call_fds[1]);

	return rc;
}

static struct virtqueue *vu_setup_vq(struct virtio_device *vdev,
				     unsigned index, vq_callback_t *callback,
				     const char *name, bool ctx)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);
	struct platform_device *pdev = vu_dev->pdev;
	struct virtio_uml_vq_info *info;
	struct virtqueue *vq;
	int num = MAX_SUPPORTED_QUEUE_SIZE;
	int rc;

	info = kzalloc(sizeof(*info), GFP_KERNEL);
	if (!info) {
		rc = -ENOMEM;
		goto error_kzalloc;
	}
	snprintf(info->name, sizeof(info->name), "%s.%d-%s", pdev->name,
		 pdev->id, name);

	vq = vring_create_virtqueue(index, num, PAGE_SIZE, vdev, true, true,
				    ctx, vu_notify, callback, info->name);
	if (!vq) {
		rc = -ENOMEM;
		goto error_create;
	}
	vq->priv = info;
	vq->num_max = num;
	num = virtqueue_get_vring_size(vq);

	if (vu_dev->protocol_features &
			BIT_ULL(VHOST_USER_PROTOCOL_F_INBAND_NOTIFICATIONS)) {
		info->kick_fd = -1;
	} else {
		rc = os_eventfd(0, 0);
		if (rc < 0)
			goto error_kick;
		info->kick_fd = rc;
	}

	rc = vu_setup_vq_call_fd(vu_dev, vq);
	if (rc)
		goto error_call;

	rc = vhost_user_set_vring_num(vu_dev, index, num);
	if (rc)
		goto error_setup;

	rc = vhost_user_set_vring_base(vu_dev, index, 0);
	if (rc)
		goto error_setup;

	rc = vhost_user_set_vring_addr(vu_dev, index,
				       virtqueue_get_desc_addr(vq),
				       virtqueue_get_used_addr(vq),
				       virtqueue_get_avail_addr(vq),
				       (u64) -1);
	if (rc)
		goto error_setup;

	return vq;

error_setup:
	if (info->call_fd >= 0) {
		um_free_irq(vu_dev->irq, vq);
		os_close_file(info->call_fd);
	}
error_call:
	if (info->kick_fd >= 0)
		os_close_file(info->kick_fd);
error_kick:
	vring_del_virtqueue(vq);
error_create:
	kfree(info);
error_kzalloc:
	return ERR_PTR(rc);
}

static int vu_find_vqs(struct virtio_device *vdev, unsigned nvqs,
		       struct virtqueue *vqs[], vq_callback_t *callbacks[],
		       const char * const names[], const bool *ctx,
		       struct irq_affinity *desc)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);
	int i, queue_idx = 0, rc;
	struct virtqueue *vq;

	/* not supported for now */
	if (WARN_ON(nvqs > 64))
		return -EINVAL;

	rc = vhost_user_set_mem_table(vu_dev);
	if (rc)
		return rc;

	for (i = 0; i < nvqs; ++i) {
		if (!names[i]) {
			vqs[i] = NULL;
			continue;
		}

		vqs[i] = vu_setup_vq(vdev, queue_idx++, callbacks[i], names[i],
				     ctx ? ctx[i] : false);
		if (IS_ERR(vqs[i])) {
			rc = PTR_ERR(vqs[i]);
			goto error_setup;
		}
	}

	list_for_each_entry(vq, &vdev->vqs, list) {
		struct virtio_uml_vq_info *info = vq->priv;

		if (info->kick_fd >= 0) {
			rc = vhost_user_set_vring_kick(vu_dev, vq->index,
						       info->kick_fd);
			if (rc)
				goto error_setup;
		}

		rc = vhost_user_set_vring_enable(vu_dev, vq->index, true);
		if (rc)
			goto error_setup;
	}

	return 0;

error_setup:
	vu_del_vqs(vdev);
	return rc;
}

static u64 vu_get_features(struct virtio_device *vdev)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	return vu_dev->features;
}

static int vu_finalize_features(struct virtio_device *vdev)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);
	u64 supported = vdev->features & VHOST_USER_SUPPORTED_F;

	vring_transport_features(vdev);
	vu_dev->features = vdev->features | supported;

	return vhost_user_set_features(vu_dev, vu_dev->features);
}

static const char *vu_bus_name(struct virtio_device *vdev)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	return vu_dev->pdev->name;
}

static const struct virtio_config_ops virtio_uml_config_ops = {
	.get = vu_get,
	.set = vu_set,
	.get_status = vu_get_status,
	.set_status = vu_set_status,
	.reset = vu_reset,
	.find_vqs = vu_find_vqs,
	.del_vqs = vu_del_vqs,
	.get_features = vu_get_features,
	.finalize_features = vu_finalize_features,
	.bus_name = vu_bus_name,
};

static void virtio_uml_release_dev(struct device *d)
{
	struct virtio_device *vdev =
			container_of(d, struct virtio_device, dev);
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	time_travel_propagate_time();

	/* might not have been opened due to not negotiating the feature */
	if (vu_dev->req_fd >= 0) {
		um_free_irq(vu_dev->irq, vu_dev);
		os_close_file(vu_dev->req_fd);
	}

	os_close_file(vu_dev->sock);
	kfree(vu_dev);
}

void virtio_uml_set_no_vq_suspend(struct virtio_device *vdev,
				  bool no_vq_suspend)
{
	struct virtio_uml_device *vu_dev = to_virtio_uml_device(vdev);

	if (WARN_ON(vdev->config != &virtio_uml_config_ops))
		return;

	vu_dev->no_vq_suspend = no_vq_suspend;
	dev_info(&vdev->dev, "%sabled VQ suspend\n",
		 no_vq_suspend ? "dis" : "en");
}

static void vu_of_conn_broken(struct work_struct *wk)
{
	struct virtio_uml_platform_data *pdata;
	struct virtio_uml_device *vu_dev;

	pdata = container_of(wk, struct virtio_uml_platform_data, conn_broken_wk);

	vu_dev = platform_get_drvdata(pdata->pdev);

	virtio_break_device(&vu_dev->vdev);

	/*
	 * We can't remove the device from the devicetree so the only thing we
	 * can do is warn.
	 */
	WARN_ON(1);
}

/* Platform device */

static struct virtio_uml_platform_data *
virtio_uml_create_pdata(struct platform_device *pdev)
{
	struct device_node *np = pdev->dev.of_node;
	struct virtio_uml_platform_data *pdata;
	int ret;

	if (!np)
		return ERR_PTR(-EINVAL);

	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
	if (!pdata)
		return ERR_PTR(-ENOMEM);

	INIT_WORK(&pdata->conn_broken_wk, vu_of_conn_broken);
	pdata->pdev = pdev;

	ret = of_property_read_string(np, "socket-path", &pdata->socket_path);
	if (ret)
		return ERR_PTR(ret);

	ret = of_property_read_u32(np, "virtio-device-id",
				   &pdata->virtio_device_id);
	if (ret)
		return ERR_PTR(ret);

	return pdata;
}

static int virtio_uml_probe(struct platform_device *pdev)
{
	struct virtio_uml_platform_data *pdata = pdev->dev.platform_data;
	struct virtio_uml_device *vu_dev;
	int rc;

	if (!pdata) {
		pdata = virtio_uml_create_pdata(pdev);
		if (IS_ERR(pdata))
			return PTR_ERR(pdata);
	}

	vu_dev = kzalloc(sizeof(*vu_dev), GFP_KERNEL);
	if (!vu_dev)
		return -ENOMEM;

	vu_dev->pdata = pdata;
	vu_dev->vdev.dev.parent = &pdev->dev;
	vu_dev->vdev.dev.release = virtio_uml_release_dev;
	vu_dev->vdev.config = &virtio_uml_config_ops;
	vu_dev->vdev.id.device = pdata->virtio_device_id;
	vu_dev->vdev.id.vendor = VIRTIO_DEV_ANY_ID;
	vu_dev->pdev = pdev;
	vu_dev->req_fd = -1;

	time_travel_propagate_time();

	do {
		rc = os_connect_socket(pdata->socket_path);
	} while (rc == -EINTR);
	if (rc < 0)
		goto error_free;
	vu_dev->sock = rc;

	spin_lock_init(&vu_dev->sock_lock);

	rc = vhost_user_init(vu_dev);
	if (rc)
		goto error_init;

	platform_set_drvdata(pdev, vu_dev);

	device_set_wakeup_capable(&vu_dev->vdev.dev, true);

	rc = register_virtio_device(&vu_dev->vdev);
	if (rc)
		put_device(&vu_dev->vdev.dev);
	vu_dev->registered = 1;
	return rc;

error_init:
	os_close_file(vu_dev->sock);
error_free:
	kfree(vu_dev);
	return rc;
}

static void virtio_uml_remove(struct platform_device *pdev)
{
	struct virtio_uml_device *vu_dev = platform_get_drvdata(pdev);

	unregister_virtio_device(&vu_dev->vdev);
}

/* Command line device list */

static void vu_cmdline_release_dev(struct device *d)
{
}

static struct device vu_cmdline_parent = {
	.init_name = "virtio-uml-cmdline",
	.release = vu_cmdline_release_dev,
};

static bool vu_cmdline_parent_registered;
static int vu_cmdline_id;

static int vu_unregister_cmdline_device(struct device *dev, void *data)
{
	struct platform_device *pdev = to_platform_device(dev);
	struct virtio_uml_platform_data *pdata = pdev->dev.platform_data;

	kfree(pdata->socket_path);
	platform_device_unregister(pdev);
	return 0;
}

static void vu_conn_broken(struct work_struct *wk)
{
	struct virtio_uml_platform_data *pdata;
	struct virtio_uml_device *vu_dev;

	pdata = container_of(wk, struct virtio_uml_platform_data, conn_broken_wk);

	vu_dev = platform_get_drvdata(pdata->pdev);

	virtio_break_device(&vu_dev->vdev);

	vu_unregister_cmdline_device(&pdata->pdev->dev, NULL);
}

static int vu_cmdline_set(const char *device, const struct kernel_param *kp)
{
	const char *ids = strchr(device, ':');
	unsigned int virtio_device_id;
	int processed, consumed, err;
	char *socket_path;
	struct virtio_uml_platform_data pdata, *ppdata;
	struct platform_device *pdev;

	if (!ids || ids == device)
		return -EINVAL;

	processed = sscanf(ids, ":%u%n:%d%n",
			   &virtio_device_id, &consumed,
			   &vu_cmdline_id, &consumed);

	if (processed < 1 || ids[consumed])
		return -EINVAL;

	if (!vu_cmdline_parent_registered) {
		err = device_register(&vu_cmdline_parent);
		if (err) {
			pr_err("Failed to register parent device!\n");
			put_device(&vu_cmdline_parent);
			return err;
		}
		vu_cmdline_parent_registered = true;
	}

	socket_path = kmemdup_nul(device, ids - device, GFP_KERNEL);
	if (!socket_path)
		return -ENOMEM;

	pdata.virtio_device_id = (u32) virtio_device_id;
	pdata.socket_path = socket_path;

	pr_info("Registering device virtio-uml.%d id=%d at %s\n",
		vu_cmdline_id, virtio_device_id, socket_path);

	pdev = platform_device_register_data(&vu_cmdline_parent, "virtio-uml",
					     vu_cmdline_id++, &pdata,
					     sizeof(pdata));
	err = PTR_ERR_OR_ZERO(pdev);
	if (err)
		goto free;

	ppdata = pdev->dev.platform_data;
	ppdata->pdev = pdev;
	INIT_WORK(&ppdata->conn_broken_wk, vu_conn_broken);

	return 0;

free:
	kfree(socket_path);
	return err;
}

static int vu_cmdline_get_device(struct device *dev, void *data)
{
	struct platform_device *pdev = to_platform_device(dev);
	struct virtio_uml_platform_data *pdata = pdev->dev.platform_data;
	char *buffer = data;
	unsigned int len = strlen(buffer);

	snprintf(buffer + len, PAGE_SIZE - len, "%s:%d:%d\n",
		 pdata->socket_path, pdata->virtio_device_id, pdev->id);
	return 0;
}

static int vu_cmdline_get(char *buffer, const struct kernel_param *kp)
{
	buffer[0] = '\0';
	if (vu_cmdline_parent_registered)
		device_for_each_child(&vu_cmdline_parent, buffer,
				      vu_cmdline_get_device);
	return strlen(buffer) + 1;
}

static const struct kernel_param_ops vu_cmdline_param_ops = {
	.set = vu_cmdline_set,
	.get = vu_cmdline_get,
};

device_param_cb(device, &vu_cmdline_param_ops, NULL, S_IRUSR);
__uml_help(vu_cmdline_param_ops,
"virtio_uml.device=<socket>:<virtio_id>[:<platform_id>]\n"
"    Configure a virtio device over a vhost-user socket.\n"
"    See virtio_ids.h for a list of possible virtio device id values.\n"
"    Optionally use a specific platform_device id.\n\n"
);


static void vu_unregister_cmdline_devices(void)
{
	if (vu_cmdline_parent_registered) {
		device_for_each_child(&vu_cmdline_parent, NULL,
				      vu_unregister_cmdline_device);
		device_unregister(&vu_cmdline_parent);
		vu_cmdline_parent_registered = false;
	}
}

/* Platform driver */

static const struct of_device_id virtio_uml_match[] = {
	{ .compatible = "virtio,uml", },
	{ }
};
MODULE_DEVICE_TABLE(of, virtio_uml_match);

static int virtio_uml_suspend(struct platform_device *pdev, pm_message_t state)
{
	struct virtio_uml_device *vu_dev = platform_get_drvdata(pdev);

	if (!vu_dev->no_vq_suspend) {
		struct virtqueue *vq;

		virtio_device_for_each_vq((&vu_dev->vdev), vq) {
			struct virtio_uml_vq_info *info = vq->priv;

			info->suspended = true;
			vhost_user_set_vring_enable(vu_dev, vq->index, false);
		}
	}

	if (!device_may_wakeup(&vu_dev->vdev.dev)) {
		vu_dev->suspended = true;
		return 0;
	}

	return irq_set_irq_wake(vu_dev->irq, 1);
}

static int virtio_uml_resume(struct platform_device *pdev)
{
	struct virtio_uml_device *vu_dev = platform_get_drvdata(pdev);

	if (!vu_dev->no_vq_suspend) {
		struct virtqueue *vq;

		virtio_device_for_each_vq((&vu_dev->vdev), vq) {
			struct virtio_uml_vq_info *info = vq->priv;

			info->suspended = false;
			vhost_user_set_vring_enable(vu_dev, vq->index, true);
		}
	}

	vu_dev->suspended = false;

	if (!device_may_wakeup(&vu_dev->vdev.dev))
		return 0;

	return irq_set_irq_wake(vu_dev->irq, 0);
}

static struct platform_driver virtio_uml_driver = {
	.probe = virtio_uml_probe,
	.remove_new = virtio_uml_remove,
	.driver = {
		.name = "virtio-uml",
		.of_match_table = virtio_uml_match,
	},
	.suspend = virtio_uml_suspend,
	.resume = virtio_uml_resume,
};

static int __init virtio_uml_init(void)
{
	return platform_driver_register(&virtio_uml_driver);
}

static void __exit virtio_uml_exit(void)
{
	platform_driver_unregister(&virtio_uml_driver);
	vu_unregister_cmdline_devices();
}

module_init(virtio_uml_init);
module_exit(virtio_uml_exit);
__uml_exitcall(virtio_uml_exit);

MODULE_DESCRIPTION("UML driver for vhost-user virtio devices");
MODULE_LICENSE("GPL");