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
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
// Copyright 2017 The ChromiumOS Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

use std::collections::BTreeMap;
use std::convert::TryFrom;
use std::fs::OpenOptions;
use std::ops::RangeInclusive;
use std::os::unix::net::UnixStream;
use std::path::Path;
use std::path::PathBuf;
use std::str;
use std::sync::Arc;

use anyhow::anyhow;
use anyhow::bail;
use anyhow::Context;
use anyhow::Result;
use arch::VirtioDeviceStub;
use base::ReadNotifier;
use base::*;
use devices::serial_device::SerialHardware;
use devices::serial_device::SerialParameters;
use devices::serial_device::SerialType;
use devices::vfio::VfioCommonSetup;
use devices::vfio::VfioCommonTrait;
use devices::virtio;
use devices::virtio::block::DiskOption;
use devices::virtio::console::asynchronous::AsyncConsole;
#[cfg(any(feature = "video-decoder", feature = "video-encoder"))]
use devices::virtio::device_constants::video::VideoBackendType;
use devices::virtio::device_constants::video::VideoDeviceType;
use devices::virtio::ipc_memory_mapper::create_ipc_mapper;
use devices::virtio::ipc_memory_mapper::CreateIpcMapperRet;
use devices::virtio::memory_mapper::BasicMemoryMapper;
use devices::virtio::memory_mapper::MemoryMapperTrait;
use devices::virtio::scsi::ScsiOption;
#[cfg(feature = "audio")]
use devices::virtio::snd::parameters::Parameters as SndParameters;
use devices::virtio::vfio_wrapper::VfioWrapper;
use devices::virtio::vhost::user::vmm::VhostUserVirtioDevice;
#[cfg(feature = "net")]
use devices::virtio::vhost::user::NetBackend;
use devices::virtio::vhost::user::VhostUserDevice;
use devices::virtio::vhost::user::VhostUserVsockDevice;
use devices::virtio::vsock::VsockConfig;
#[cfg(feature = "balloon")]
use devices::virtio::BalloonMode;
use devices::virtio::Console;
#[cfg(feature = "net")]
use devices::virtio::NetError;
#[cfg(feature = "net")]
use devices::virtio::NetParameters;
#[cfg(feature = "net")]
use devices::virtio::NetParametersMode;
use devices::virtio::VirtioDevice;
use devices::virtio::VirtioDeviceType;
use devices::BusDeviceObj;
use devices::IommuDevType;
use devices::PciAddress;
use devices::PciDevice;
use devices::VfioDevice;
use devices::VfioDeviceType;
use devices::VfioPciDevice;
use devices::VfioPlatformDevice;
#[cfg(feature = "vtpm")]
use devices::VtpmProxy;
use hypervisor::ProtectionType;
use hypervisor::Vm;
use jail::*;
use minijail::Minijail;
#[cfg(feature = "net")]
use net_util::sys::unix::Tap;
#[cfg(feature = "net")]
use net_util::MacAddress;
#[cfg(feature = "net")]
use net_util::TapTCommon;
use resources::Alloc;
use resources::AllocOptions;
use resources::SystemAllocator;
use sync::Mutex;
use vm_control::api::VmMemoryClient;
use vm_memory::GuestAddress;

use crate::crosvm::config::TouchDeviceOption;
use crate::crosvm::config::VhostUserFsOption;
use crate::crosvm::config::VhostUserOption;

pub enum TaggedControlTube {
    Fs(Tube),
    Vm(Tube),
    VmMsync(Tube),
}

impl AsRef<Tube> for TaggedControlTube {
    fn as_ref(&self) -> &Tube {
        use self::TaggedControlTube::*;
        match &self {
            Fs(tube) | Vm(tube) | VmMsync(tube) => tube,
        }
    }
}

impl AsRawDescriptor for TaggedControlTube {
    fn as_raw_descriptor(&self) -> RawDescriptor {
        self.as_ref().as_raw_descriptor()
    }
}

impl ReadNotifier for TaggedControlTube {
    fn get_read_notifier(&self) -> &dyn AsRawDescriptor {
        self.as_ref().get_read_notifier()
    }
}

#[derive(serde::Serialize, serde::Deserialize)]
pub struct VmMemoryTube {
    pub tube: Tube,
    /// See devices::virtio::VirtioDevice.expose_shared_memory_region_with_viommu
    pub expose_with_viommu: bool,
}

impl AsRef<Tube> for VmMemoryTube {
    fn as_ref(&self) -> &Tube {
        &self.tube
    }
}

impl AsRawDescriptor for VmMemoryTube {
    fn as_raw_descriptor(&self) -> RawDescriptor {
        self.as_ref().as_raw_descriptor()
    }
}

impl ReadNotifier for VmMemoryTube {
    fn get_read_notifier(&self) -> &dyn AsRawDescriptor {
        self.as_ref().get_read_notifier()
    }
}

pub trait IntoUnixStream {
    fn into_unix_stream(self) -> Result<UnixStream>;
}

impl<'a> IntoUnixStream for &'a Path {
    fn into_unix_stream(self) -> Result<UnixStream> {
        if let Some(fd) = safe_descriptor_from_path(self).context("failed to open event device")? {
            Ok(fd.into())
        } else {
            UnixStream::connect(self).context("failed to open event device")
        }
    }
}

impl<'a> IntoUnixStream for &'a PathBuf {
    fn into_unix_stream(self) -> Result<UnixStream> {
        self.as_path().into_unix_stream()
    }
}

impl IntoUnixStream for UnixStream {
    fn into_unix_stream(self) -> Result<UnixStream> {
        Ok(self)
    }
}

pub type DeviceResult<T = VirtioDeviceStub> = Result<T>;

/// A trait for spawning virtio device instances and jails from their configuration structure.
///
/// Implementors become able to create virtio devices and jails following their own configuration.
/// This trait also provides a few convenience methods for e.g. creating a virtio device and jail
/// at once.
pub trait VirtioDeviceBuilder: Sized {
    /// Base name of the device, as it will appear in logs.
    const NAME: &'static str;

    /// Create a regular virtio device from the configuration and `protection_type` setting.
    fn create_virtio_device(
        self,
        protection_type: ProtectionType,
    ) -> anyhow::Result<Box<dyn VirtioDevice>>;

    /// Create a device suitable for being run as a vhost-user instance.
    ///
    /// It is ok to leave this method unimplemented if the device is not intended to be used with
    /// vhost-user.
    fn create_vhost_user_device(
        self,
        _keep_rds: &mut Vec<RawDescriptor>,
    ) -> anyhow::Result<Box<dyn VhostUserDevice>> {
        unimplemented!()
    }

    /// Create a jail that is suitable to run a device.
    ///
    /// The default implementation creates a simple jail with a seccomp policy derived from the
    /// base name of the device.
    fn create_jail(
        &self,
        jail_config: &Option<JailConfig>,
        virtio_transport: VirtioDeviceType,
    ) -> anyhow::Result<Option<Minijail>> {
        simple_jail(
            jail_config,
            &virtio_transport.seccomp_policy_file(Self::NAME),
        )
    }

    /// Helper method to return a `VirtioDeviceStub` filled using `create_virtio_device` and
    /// `create_jail`.
    ///
    /// This helper should cover the needs of most devices when run as regular virtio devices.
    fn create_virtio_device_and_jail(
        self,
        protection_type: ProtectionType,
        jail_config: &Option<JailConfig>,
    ) -> DeviceResult {
        let jail = self.create_jail(jail_config, VirtioDeviceType::Regular)?;
        let dev = self.create_virtio_device(protection_type)?;
        Ok(VirtioDeviceStub { dev, jail })
    }
}

/// A one-shot configuration structure for implementing `VirtioDeviceBuilder`. We cannot do it on
/// `DiskOption` directly because disk devices can be passed an optional control tube.
pub struct DiskConfig<'a> {
    /// Options for disk creation.
    disk: &'a DiskOption,
    /// Optional control tube for the device.
    device_tube: Option<Tube>,
}

impl<'a> DiskConfig<'a> {
    pub fn new(disk: &'a DiskOption, device_tube: Option<Tube>) -> Self {
        Self { disk, device_tube }
    }
}

impl<'a> VirtioDeviceBuilder for DiskConfig<'a> {
    const NAME: &'static str = "block";

    fn create_virtio_device(
        self,
        protection_type: ProtectionType,
    ) -> anyhow::Result<Box<dyn VirtioDevice>> {
        info!(
            "Trying to attach block device: {}",
            self.disk.path.display(),
        );
        let disk_image = self.disk.open()?;
        let base_features = virtio::base_features(protection_type);
        Ok(Box::new(
            virtio::BlockAsync::new(
                base_features,
                disk_image,
                self.disk,
                self.device_tube,
                None,
                None,
            )
            .context("failed to create block device")?,
        ))
    }

    fn create_vhost_user_device(
        self,
        keep_rds: &mut Vec<RawDescriptor>,
    ) -> anyhow::Result<Box<dyn VhostUserDevice>> {
        let disk = self.disk;
        let disk_image = disk.open()?;
        let base_features = virtio::base_features(ProtectionType::Unprotected);

        let block = Box::new(
            virtio::BlockAsync::new(
                base_features,
                disk_image,
                disk,
                self.device_tube,
                None,
                None,
            )
            .context("failed to create block device")?,
        );
        keep_rds.extend(block.keep_rds());

        Ok(block)
    }
}

impl<'a> VirtioDeviceBuilder for &'a ScsiOption {
    const NAME: &'static str = "scsi";

    fn create_virtio_device(
        self,
        protection_type: ProtectionType,
    ) -> anyhow::Result<Box<dyn VirtioDevice>> {
        let base_features = virtio::base_features(protection_type);
        info!("Trying to attach scsi device: {}", self.path.display());
        let disk_image = self.open()?;
        Ok(Box::new(
            virtio::ScsiDevice::new(disk_image, base_features, self.block_size, self.read_only)
                .context("failed to create scsi device")?,
        ))
    }
}

fn vhost_user_connection(path: &Path) -> Result<UnixStream> {
    UnixStream::connect(path).with_context(|| {
        format!(
            "failed to connect to vhost-user socket path {}",
            path.display()
        )
    })
}

pub fn create_vhost_user_block_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_block(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user block device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_console_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_console(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user console device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_fs_device(
    protection_type: ProtectionType,
    option: &VhostUserFsOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_fs(
        virtio::base_features(protection_type),
        vhost_user_connection(&option.socket)?,
        option.max_queue_size,
        &option.tag,
    )
    .context("failed to set up vhost-user fs device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_mac80211_hwsim_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_mac80211_hwsim(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user mac80211_hwsim device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_snd_device(
    protection_type: ProtectionType,
    option: &VhostUserOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_snd(
        virtio::base_features(protection_type),
        vhost_user_connection(&option.socket)?,
        option.max_queue_size,
    )
    .context("failed to set up vhost-user snd device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_gpu_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    // The crosvm gpu device expects us to connect the tube before it will accept a vhost-user
    // connection.
    let dev = VhostUserVirtioDevice::new_gpu(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user gpu device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_rng_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
) -> DeviceResult {
    let dev =
        virtio::Rng::new(virtio::base_features(protection_type)).context("failed to set up rng")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "rng_device")?,
    })
}

#[cfg(feature = "audio")]
pub fn create_virtio_snd_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    snd_params: SndParameters,
) -> DeviceResult {
    let backend = snd_params.backend;
    let dev = virtio::snd::common_backend::VirtioSnd::new(
        virtio::base_features(protection_type),
        snd_params,
    )
    .context("failed to create cras sound device")?;

    use virtio::snd::parameters::StreamSourceBackend as Backend;

    let policy = match backend {
        Backend::NULL | Backend::FILE => "snd_null_device",
        #[cfg(feature = "audio_cras")]
        Backend::Sys(virtio::snd::sys::StreamSourceBackend::CRAS) => "snd_cras_device",
        #[cfg(not(feature = "audio_cras"))]
        _ => unreachable!(),
    };

    let jail = if let Some(jail_config) = jail_config {
        let mut config = SandboxConfig::new(jail_config, policy);
        #[cfg(feature = "audio_cras")]
        if backend == Backend::Sys(virtio::snd::sys::StreamSourceBackend::CRAS) {
            config.bind_mounts = true;
        }
        // TODO(b/267574679): running as current_user may not be required for snd device.
        config.run_as = RunAsUser::CurrentUser;
        #[allow(unused_mut)]
        let mut jail =
            create_sandbox_minijail(&jail_config.pivot_root, MAX_OPEN_FILES_DEFAULT, &config)?;
        #[cfg(feature = "audio_cras")]
        if backend == Backend::Sys(virtio::snd::sys::StreamSourceBackend::CRAS) {
            let run_cras_path = Path::new("/run/cras");
            jail.mount_bind(run_cras_path, run_cras_path, true)?;
        }
        Some(jail)
    } else {
        None
    };

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail,
    })
}

#[cfg(feature = "vtpm")]
pub fn create_vtpm_proxy_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
) -> DeviceResult {
    let jail = if let Some(jail_config) = jail_config {
        let mut config = SandboxConfig::new(jail_config, "vtpm_proxy_device");
        config.bind_mounts = true;
        let mut jail =
            create_sandbox_minijail(&jail_config.pivot_root, MAX_OPEN_FILES_DEFAULT, &config)?;
        let system_bus_socket_path = Path::new("/run/dbus/system_bus_socket");
        jail.mount_bind(system_bus_socket_path, system_bus_socket_path, true)?;
        Some(jail)
    } else {
        None
    };

    let backend = VtpmProxy::new();
    let dev = virtio::Tpm::new(Box::new(backend), virtio::base_features(protection_type));

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail,
    })
}

pub fn create_single_touch_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    single_touch_spec: &TouchDeviceOption,
    idx: u32,
) -> DeviceResult {
    let socket = single_touch_spec
        .get_path()
        .into_unix_stream()
        .context("failed configuring virtio single touch")?;

    let (width, height) = single_touch_spec.get_size();
    let name = single_touch_spec.get_name();
    let dev = virtio::input::new_single_touch(
        idx,
        socket,
        width,
        height,
        name,
        virtio::base_features(protection_type),
    )
    .context("failed to set up input device")?;
    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_multi_touch_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    multi_touch_spec: &TouchDeviceOption,
    idx: u32,
) -> DeviceResult {
    let socket = multi_touch_spec
        .get_path()
        .into_unix_stream()
        .context("failed configuring virtio multi touch")?;

    let (width, height) = multi_touch_spec.get_size();
    let name = multi_touch_spec.get_name();
    let dev = virtio::input::new_multi_touch(
        idx,
        socket,
        width,
        height,
        name,
        virtio::base_features(protection_type),
    )
    .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_trackpad_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    trackpad_spec: &TouchDeviceOption,
    idx: u32,
) -> DeviceResult {
    let socket = trackpad_spec
        .get_path()
        .into_unix_stream()
        .context("failed configuring virtio trackpad")?;

    let (width, height) = trackpad_spec.get_size();
    let name = trackpad_spec.get_name();
    let dev = virtio::input::new_trackpad(
        idx,
        socket,
        width,
        height,
        name,
        virtio::base_features(protection_type),
    )
    .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_mouse_device<T: IntoUnixStream>(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    mouse_socket: T,
    idx: u32,
) -> DeviceResult {
    let socket = mouse_socket
        .into_unix_stream()
        .context("failed configuring virtio mouse")?;

    let dev = virtio::input::new_mouse(idx, socket, virtio::base_features(protection_type))
        .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_keyboard_device<T: IntoUnixStream>(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    keyboard_socket: T,
    idx: u32,
) -> DeviceResult {
    let socket = keyboard_socket
        .into_unix_stream()
        .context("failed configuring virtio keyboard")?;

    let dev = virtio::input::new_keyboard(idx, socket, virtio::base_features(protection_type))
        .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_switches_device<T: IntoUnixStream>(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    switches_socket: T,
    idx: u32,
) -> DeviceResult {
    let socket = switches_socket
        .into_unix_stream()
        .context("failed configuring virtio switches")?;

    let dev = virtio::input::new_switches(idx, socket, virtio::base_features(protection_type))
        .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_rotary_device<T: IntoUnixStream>(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    rotary_socket: T,
    idx: u32,
) -> DeviceResult {
    let socket = rotary_socket
        .into_unix_stream()
        .context("failed configuring virtio rotary")?;

    let dev = virtio::input::new_rotary(idx, socket, virtio::base_features(protection_type))
        .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

pub fn create_vinput_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    dev_path: &Path,
) -> DeviceResult {
    let dev_file = OpenOptions::new()
        .read(true)
        .write(true)
        .open(dev_path)
        .with_context(|| format!("failed to open vinput device {}", dev_path.display()))?;

    let dev = virtio::input::new_evdev(dev_file, virtio::base_features(protection_type))
        .context("failed to set up input device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "input_device")?,
    })
}

#[cfg(feature = "balloon")]
pub fn create_balloon_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    mode: BalloonMode,
    tube: Tube,
    inflate_tube: Option<Tube>,
    init_balloon_size: u64,
    enabled_features: u64,
    #[cfg(feature = "registered_events")] registered_evt_q: Option<SendTube>,
    ws_num_bins: u8,
) -> DeviceResult {
    let dev = virtio::Balloon::new(
        virtio::base_features(protection_type),
        tube,
        inflate_tube,
        init_balloon_size,
        mode,
        enabled_features,
        #[cfg(feature = "registered_events")]
        registered_evt_q,
        ws_num_bins,
    )
    .context("failed to create balloon")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "balloon_device")?,
    })
}

#[cfg(feature = "net")]
impl VirtioDeviceBuilder for &NetParameters {
    const NAME: &'static str = "net";

    fn create_virtio_device(
        self,
        protection_type: ProtectionType,
    ) -> anyhow::Result<Box<dyn VirtioDevice>> {
        let vq_pairs = self.vq_pairs.unwrap_or(1);
        let multi_vq = vq_pairs > 1 && self.vhost_net.is_none();

        let features = virtio::base_features(protection_type);
        let (tap, mac) = create_tap_for_net_device(&self.mode, multi_vq)?;

        Ok(if let Some(vhost_net) = &self.vhost_net {
            Box::new(
                virtio::vhost::Net::<_, vhost::Net<_>>::new(
                    &vhost_net.device,
                    features,
                    tap,
                    mac,
                    self.packed_queue,
                )
                .context("failed to set up virtio-vhost networking")?,
            ) as Box<dyn VirtioDevice>
        } else {
            Box::new(
                virtio::Net::new(features, tap, vq_pairs, mac, self.packed_queue)
                    .context("failed to set up virtio networking")?,
            ) as Box<dyn VirtioDevice>
        })
    }

    fn create_jail(
        &self,
        jail_config: &Option<JailConfig>,
        virtio_transport: VirtioDeviceType,
    ) -> anyhow::Result<Option<Minijail>> {
        let policy = if self.vhost_net.is_some() {
            "vhost_net"
        } else {
            "net"
        };

        simple_jail(jail_config, &virtio_transport.seccomp_policy_file(policy))
    }

    fn create_vhost_user_device(
        self,
        keep_rds: &mut Vec<RawDescriptor>,
    ) -> anyhow::Result<Box<dyn VhostUserDevice>> {
        let vq_pairs = self.vq_pairs.unwrap_or(1);
        let multi_vq = vq_pairs > 1 && self.vhost_net.is_none();
        let (tap, _mac) = create_tap_for_net_device(&self.mode, multi_vq)?;

        let backend = NetBackend::new(tap)?;

        keep_rds.extend(backend.as_raw_descriptors());

        Ok(Box::new(backend))
    }
}

/// Create a new tap interface based on NetParametersMode.
#[cfg(feature = "net")]
fn create_tap_for_net_device(
    mode: &NetParametersMode,
    multi_vq: bool,
) -> DeviceResult<(Tap, Option<MacAddress>)> {
    match mode {
        NetParametersMode::TapName { tap_name, mac } => {
            let tap = Tap::new_with_name(tap_name.as_bytes(), true, multi_vq)
                .map_err(NetError::TapOpen)?;
            Ok((tap, *mac))
        }
        NetParametersMode::TapFd { tap_fd, mac } => {
            // Safe because we ensure that we get a unique handle to the fd.
            let tap = unsafe {
                Tap::from_raw_descriptor(
                    validate_raw_descriptor(*tap_fd)
                        .context("failed to validate tap descriptor")?,
                )
                .context("failed to create tap device")?
            };
            Ok((tap, *mac))
        }
        NetParametersMode::RawConfig {
            host_ip,
            netmask,
            mac,
        } => {
            let tap = Tap::new(true, multi_vq).map_err(NetError::TapOpen)?;
            tap.set_ip_addr(*host_ip).map_err(NetError::TapSetIp)?;
            tap.set_netmask(*netmask).map_err(NetError::TapSetNetmask)?;
            tap.set_mac_address(*mac)
                .map_err(NetError::TapSetMacAddress)?;
            tap.enable().map_err(NetError::TapEnable)?;
            Ok((tap, None))
        }
    }
}

pub fn create_vhost_user_net_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_net(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user net device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_vsock_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_vsock(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user vsock device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_vhost_user_wl_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
) -> DeviceResult {
    // The crosvm wl device expects us to connect the tube before it will accept a vhost-user
    // connection.
    let dev = VhostUserVirtioDevice::new_wl(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
    )
    .context("failed to set up vhost-user wl device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

pub fn create_wayland_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    wayland_socket_paths: &BTreeMap<String, PathBuf>,
    resource_bridge: Option<Tube>,
) -> DeviceResult {
    let wayland_socket_dirs = wayland_socket_paths
        .iter()
        .map(|(_name, path)| path.parent())
        .collect::<Option<Vec<_>>>()
        .ok_or_else(|| anyhow!("wayland socket path has no parent or file name"))?;

    let features = virtio::base_features(protection_type);
    let dev = virtio::Wl::new(features, wayland_socket_paths.clone(), resource_bridge)
        .context("failed to create wayland device")?;

    let jail = if let Some(jail_config) = jail_config {
        let mut config = SandboxConfig::new(jail_config, "wl_device");
        config.bind_mounts = true;
        let mut jail = create_gpu_minijail(
            &jail_config.pivot_root,
            &config,
            /* render_node_only= */ false,
        )?;
        // Bind mount the wayland socket's directory into jail's root. This is necessary since
        // each new wayland context must open() the socket. If the wayland socket is ever
        // destroyed and remade in the same host directory, new connections will be possible
        // without restarting the wayland device.
        for dir in &wayland_socket_dirs {
            jail.mount(dir, dir, "", (libc::MS_BIND | libc::MS_REC) as usize)?;
        }

        Some(jail)
    } else {
        None
    };

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail,
    })
}

#[cfg(any(feature = "video-decoder", feature = "video-encoder"))]
pub fn create_video_device(
    backend: VideoBackendType,
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    typ: VideoDeviceType,
    resource_bridge: Tube,
) -> DeviceResult {
    let jail = if let Some(jail_config) = jail_config {
        match typ {
            #[cfg(feature = "video-decoder")]
            VideoDeviceType::Decoder => {}
            #[cfg(feature = "video-encoder")]
            VideoDeviceType::Encoder => {}
            #[cfg(any(not(feature = "video-decoder"), not(feature = "video-encoder")))]
            // `typ` is always a VideoDeviceType enabled
            device_type => unreachable!("Not compiled with {:?} enabled", device_type),
        };
        let mut config = SandboxConfig::new(jail_config, "video_device");
        config.bind_mounts = true;
        let mut jail =
            create_sandbox_minijail(&jail_config.pivot_root, MAX_OPEN_FILES_DEFAULT, &config)?;

        let need_drm_device = match backend {
            #[cfg(any(feature = "libvda", feature = "libvda-stub"))]
            VideoBackendType::Libvda => true,
            #[cfg(any(feature = "libvda", feature = "libvda-stub"))]
            VideoBackendType::LibvdaVd => true,
            #[cfg(feature = "vaapi")]
            VideoBackendType::Vaapi => true,
            #[cfg(feature = "ffmpeg")]
            VideoBackendType::Ffmpeg => false,
        };

        if need_drm_device {
            jail_mount_bind_drm(&mut jail, /* render_node_only= */ true)?;
        }

        #[cfg(target_arch = "x86_64")]
        {
            // Device nodes used by libdrm through minigbm in libvda on AMD devices.
            let sys_dev_char_path = Path::new("/sys/dev/char");
            jail.mount_bind(sys_dev_char_path, sys_dev_char_path, false)?;
            let sys_devices_path = Path::new("/sys/devices");
            jail.mount_bind(sys_devices_path, sys_devices_path, false)?;

            // Required for loading dri libraries loaded by minigbm on AMD devices.
            jail_mount_bind_if_exists(&mut jail, &["/usr/lib64", "/usr/lib"])?;
        }

        // Device nodes required by libchrome which establishes Mojo connection in libvda.
        let dev_urandom_path = Path::new("/dev/urandom");
        jail.mount_bind(dev_urandom_path, dev_urandom_path, false)?;
        let system_bus_socket_path = Path::new("/run/dbus/system_bus_socket");
        jail.mount_bind(system_bus_socket_path, system_bus_socket_path, true)?;

        Some(jail)
    } else {
        None
    };

    Ok(VirtioDeviceStub {
        dev: Box::new(devices::virtio::VideoDevice::new(
            virtio::base_features(protection_type),
            typ,
            backend,
            Some(resource_bridge),
        )),
        jail,
    })
}

#[cfg(any(feature = "video-decoder", feature = "video-encoder"))]
pub fn register_video_device(
    backend: VideoBackendType,
    devs: &mut Vec<VirtioDeviceStub>,
    video_tube: Tube,
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    typ: VideoDeviceType,
) -> Result<()> {
    devs.push(create_video_device(
        backend,
        protection_type,
        jail_config,
        typ,
        video_tube,
    )?);
    Ok(())
}

pub fn create_vhost_user_video_device(
    protection_type: ProtectionType,
    opt: &VhostUserOption,
    device_type: VideoDeviceType,
) -> DeviceResult {
    let dev = VhostUserVirtioDevice::new_video(
        virtio::base_features(protection_type),
        vhost_user_connection(&opt.socket)?,
        opt.max_queue_size,
        device_type,
    )
    .context("failed to set up vhost-user video device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        // no sandbox here because virtqueue handling is exported to a different process.
        jail: None,
    })
}

impl VirtioDeviceBuilder for &VsockConfig {
    const NAME: &'static str = "vhost_vsock";

    fn create_virtio_device(
        self,
        protection_type: ProtectionType,
    ) -> anyhow::Result<Box<dyn VirtioDevice>> {
        let features = virtio::base_features(protection_type);

        let dev = virtio::vhost::Vsock::new(features, self)
            .context("failed to set up virtual socket device")?;

        Ok(Box::new(dev))
    }

    fn create_vhost_user_device(
        self,
        keep_rds: &mut Vec<RawDescriptor>,
    ) -> anyhow::Result<Box<dyn VhostUserDevice>> {
        let vsock_device = VhostUserVsockDevice::new(self.cid, &self.vhost_device)?;

        keep_rds.push(vsock_device.as_raw_descriptor());

        Ok(Box::new(vsock_device))
    }
}

#[cfg(any(target_arch = "arm", target_arch = "aarch64"))]
pub fn create_vhost_scmi_device(
    protected_vm: ProtectionType,
    jail_config: &Option<JailConfig>,
    vhost_scmi_dev_path: PathBuf,
) -> DeviceResult {
    let features = virtio::base_features(protected_vm);

    let dev = virtio::vhost::Scmi::new(&vhost_scmi_dev_path, features)
        .context("failed to set up vhost scmi device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "vhost_scmi_device")?,
    })
}

pub fn create_fs_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    ugid: (Option<u32>, Option<u32>),
    uid_map: &str,
    gid_map: &str,
    src: &Path,
    tag: &str,
    fs_cfg: virtio::fs::Config,
    device_tube: Tube,
) -> DeviceResult {
    let max_open_files =
        base::get_max_open_files().context("failed to get max number of open files")?;
    let j = if let Some(jail_config) = jail_config {
        let mut config = SandboxConfig::new(jail_config, "fs_device");
        config.limit_caps = false;
        config.ugid_map = Some((uid_map, gid_map));
        // We want bind mounts from the parent namespaces to propagate into the fs device's
        // namespace.
        config.remount_mode = Some(libc::MS_SLAVE);
        config.run_as = if ugid == (None, None) {
            RunAsUser::Unspecified
        } else {
            RunAsUser::Specified(ugid.0.unwrap_or(0), ugid.1.unwrap_or(0))
        };
        create_sandbox_minijail(src, max_open_files, &config)?
    } else {
        create_base_minijail(src, max_open_files)?
    };

    let features = virtio::base_features(protection_type);
    // TODO(chirantan): Use more than one worker once the kernel driver has been fixed to not panic
    // when num_queues > 1.
    let dev = virtio::fs::Fs::new(features, tag, 1, fs_cfg, device_tube)
        .context("failed to create fs device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: Some(j),
    })
}

pub fn create_9p_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    ugid: (Option<u32>, Option<u32>),
    uid_map: &str,
    gid_map: &str,
    src: &Path,
    tag: &str,
    mut p9_cfg: p9::Config,
) -> DeviceResult {
    let max_open_files =
        base::get_max_open_files().context("failed to get max number of open files")?;
    let (jail, root) = if let Some(jail_config) = jail_config {
        let mut config = SandboxConfig::new(jail_config, "9p_device");
        config.limit_caps = false;
        config.ugid_map = Some((uid_map, gid_map));
        // We want bind mounts from the parent namespaces to propagate into the 9p server's
        // namespace.
        config.remount_mode = Some(libc::MS_SLAVE);
        config.run_as = if ugid == (None, None) {
            RunAsUser::Unspecified
        } else {
            RunAsUser::Specified(ugid.0.unwrap_or(0), ugid.1.unwrap_or(0))
        };
        let jail = create_sandbox_minijail(src, max_open_files, &config)?;

        //  The shared directory becomes the root of the device's file system.
        let root = Path::new("/");
        (Some(jail), root)
    } else {
        // There's no mount namespace so we tell the server to treat the source directory as the
        // root.
        (None, src)
    };

    let features = virtio::base_features(protection_type);
    p9_cfg.root = root.into();
    let dev = virtio::P9::new(features, tag, p9_cfg).context("failed to create 9p device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail,
    })
}

pub fn create_pmem_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    vm: &mut impl Vm,
    resources: &mut SystemAllocator,
    disk: &DiskOption,
    index: usize,
    pmem_device_tube: Tube,
) -> DeviceResult {
    let fd = open_file_or_duplicate(
        &disk.path,
        OpenOptions::new().read(true).write(!disk.read_only),
    )
    .with_context(|| format!("failed to load disk image {}", disk.path.display()))?;

    let (disk_size, arena_size) = {
        let metadata = std::fs::metadata(&disk.path).with_context(|| {
            format!("failed to get disk image {} metadata", disk.path.display())
        })?;
        let disk_len = metadata.len();
        // Linux requires pmem region sizes to be 2 MiB aligned. Linux will fill any partial page
        // at the end of an mmap'd file and won't write back beyond the actual file length, but if
        // we just align the size of the file to 2 MiB then access beyond the last page of the
        // mapped file will generate SIGBUS. So use a memory mapping arena that will provide
        // padding up to 2 MiB.
        let alignment = 2 * 1024 * 1024;
        let align_adjust = if disk_len % alignment != 0 {
            alignment - (disk_len % alignment)
        } else {
            0
        };
        (
            disk_len,
            disk_len
                .checked_add(align_adjust)
                .ok_or_else(|| anyhow!("pmem device image too big"))?,
        )
    };

    let protection = {
        if disk.read_only {
            Protection::read()
        } else {
            Protection::read_write()
        }
    };

    let arena = {
        // Conversion from u64 to usize may fail on 32bit system.
        let arena_size = usize::try_from(arena_size).context("pmem device image too big")?;
        let disk_size = usize::try_from(disk_size).context("pmem device image too big")?;

        let mut arena =
            MemoryMappingArena::new(arena_size).context("failed to reserve pmem memory")?;
        arena
            .add_fd_offset_protection(0, disk_size, &fd, 0, protection)
            .context("failed to reserve pmem memory")?;

        // If the disk is not a multiple of the page size, the OS will fill the remaining part
        // of the page with zeroes. However, the anonymous mapping added below must start on a
        // page boundary, so round up the size before calculating the offset of the anon region.
        let disk_size = round_up_to_page_size(disk_size);

        if arena_size > disk_size {
            // Add an anonymous region with the same protection as the disk mapping if the arena
            // size was aligned.
            arena
                .add_anon_protection(disk_size, arena_size - disk_size, protection)
                .context("failed to reserve pmem padding")?;
        }
        arena
    };

    let mapping_address = resources
        .allocate_mmio(
            arena_size,
            Alloc::PmemDevice(index),
            format!("pmem_disk_image_{}", index),
            AllocOptions::new()
                .top_down(true)
                .prefetchable(true)
                // Linux kernel requires pmem namespaces to be 128 MiB aligned.
                .align(128 * 1024 * 1024), /* 128 MiB */
        )
        .context("failed to allocate memory for pmem device")?;

    let slot = vm
        .add_memory_region(
            GuestAddress(mapping_address),
            Box::new(arena),
            /* read_only = */ disk.read_only,
            /* log_dirty_pages = */ false,
        )
        .context("failed to add pmem device memory")?;

    let dev = virtio::Pmem::new(
        virtio::base_features(protection_type),
        fd,
        GuestAddress(mapping_address),
        slot,
        arena_size,
        pmem_device_tube,
    )
    .context("failed to create pmem device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev) as Box<dyn VirtioDevice>,
        jail: simple_jail(jail_config, "pmem_device")?,
    })
}

pub fn create_iommu_device(
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
    iova_max_addr: u64,
    endpoints: BTreeMap<u32, Arc<Mutex<Box<dyn MemoryMapperTrait>>>>,
    hp_endpoints_ranges: Vec<RangeInclusive<u32>>,
    translate_response_senders: Option<BTreeMap<u32, Tube>>,
    translate_request_rx: Option<Tube>,
    iommu_device_tube: Tube,
) -> DeviceResult {
    let dev = virtio::Iommu::new(
        virtio::base_features(protection_type),
        endpoints,
        iova_max_addr,
        hp_endpoints_ranges,
        translate_response_senders,
        translate_request_rx,
        Some(iommu_device_tube),
    )
    .context("failed to create IOMMU device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "iommu_device")?,
    })
}

fn add_bind_mounts(param: &SerialParameters, jail: &mut Minijail) -> Result<(), minijail::Error> {
    if let Some(path) = &param.path {
        if let SerialType::SystemSerialType = param.type_ {
            if let Some(parent) = path.as_path().parent() {
                if parent.exists() {
                    info!("Bind mounting dir {}", parent.display());
                    jail.mount_bind(parent, parent, true)?;
                }
            }
        }
    }
    Ok(())
}

/// For creating console virtio devices.
impl VirtioDeviceBuilder for &SerialParameters {
    const NAME: &'static str = "serial";

    fn create_virtio_device(
        self,
        protection_type: ProtectionType,
    ) -> anyhow::Result<Box<dyn VirtioDevice>> {
        let mut keep_rds = Vec::new();
        let evt = Event::new().context("failed to create event")?;

        if self.hardware == SerialHardware::LegacyVirtioConsole {
            Ok(Box::new(
                self.create_serial_device::<Console>(protection_type, &evt, &mut keep_rds)
                    .context("failed to create console device")?,
            ))
        } else {
            Ok(Box::new(
                self.create_serial_device::<AsyncConsole>(protection_type, &evt, &mut keep_rds)
                    .context("failed to create console device")?,
            ))
        }
    }

    fn create_vhost_user_device(
        self,
        keep_rds: &mut Vec<RawDescriptor>,
    ) -> anyhow::Result<Box<dyn VhostUserDevice>> {
        Ok(Box::new(virtio::vhost::user::create_vu_console_device(
            self, keep_rds,
        )?))
    }

    fn create_jail(
        &self,
        jail_config: &Option<JailConfig>,
        virtio_transport: VirtioDeviceType,
    ) -> anyhow::Result<Option<Minijail>> {
        if let Some(jail_config) = jail_config {
            let policy = virtio_transport.seccomp_policy_file("serial");
            let mut config = SandboxConfig::new(jail_config, &policy);
            config.bind_mounts = true;
            let mut jail =
                create_sandbox_minijail(&jail_config.pivot_root, MAX_OPEN_FILES_DEFAULT, &config)?;
            add_bind_mounts(self, &mut jail)
                .context("failed to add bind mounts for console device")?;
            Ok(Some(jail))
        } else {
            Ok(None)
        }
    }
}

#[cfg(feature = "audio")]
pub fn create_sound_device(
    path: &Path,
    protection_type: ProtectionType,
    jail_config: &Option<JailConfig>,
) -> DeviceResult {
    let dev = virtio::new_sound(path, virtio::base_features(protection_type))
        .context("failed to create sound device")?;

    Ok(VirtioDeviceStub {
        dev: Box::new(dev),
        jail: simple_jail(jail_config, "vios_audio_device")?,
    })
}

#[allow(clippy::large_enum_variant)]
pub enum VfioDeviceVariant {
    Pci(VfioPciDevice),
    Platform(VfioPlatformDevice),
}

pub fn create_vfio_device(
    jail_config: &Option<JailConfig>,
    vm: &impl Vm,
    resources: &mut SystemAllocator,
    irq_control_tubes: &mut Vec<Tube>,
    vm_memory_control_tubes: &mut Vec<VmMemoryTube>,
    control_tubes: &mut Vec<TaggedControlTube>,
    vfio_path: &Path,
    hotplug: bool,
    hotplug_bus: Option<u8>,
    guest_address: Option<PciAddress>,
    coiommu_endpoints: Option<&mut Vec<u16>>,
    iommu_dev: IommuDevType,
) -> DeviceResult<(VfioDeviceVariant, Option<Minijail>, Option<VfioWrapper>)> {
    let vfio_container = VfioCommonSetup::vfio_get_container(iommu_dev, Some(vfio_path))
        .context("failed to get vfio container")?;

    let (vfio_host_tube_mem, vfio_device_tube_mem) =
        Tube::pair().context("failed to create tube")?;
    vm_memory_control_tubes.push(VmMemoryTube {
        tube: vfio_host_tube_mem,
        expose_with_viommu: false,
    });

    let (vfio_host_tube_vm, vfio_device_tube_vm) = Tube::pair().context("failed to create tube")?;
    control_tubes.push(TaggedControlTube::Vm(vfio_host_tube_vm));

    let vfio_device = VfioDevice::new_passthrough(
        &vfio_path,
        vm,
        vfio_container.clone(),
        iommu_dev != IommuDevType::NoIommu,
    )
    .context("failed to create vfio device")?;

    match vfio_device.device_type() {
        VfioDeviceType::Pci => {
            let (vfio_host_tube_msi, vfio_device_tube_msi) =
                Tube::pair().context("failed to create tube")?;
            irq_control_tubes.push(vfio_host_tube_msi);

            let (vfio_host_tube_msix, vfio_device_tube_msix) =
                Tube::pair().context("failed to create tube")?;
            irq_control_tubes.push(vfio_host_tube_msix);

            let mut vfio_pci_device = VfioPciDevice::new(
                vfio_path,
                vfio_device,
                hotplug,
                hotplug_bus,
                guest_address,
                vfio_device_tube_msi,
                vfio_device_tube_msix,
                VmMemoryClient::new(vfio_device_tube_mem),
                vfio_device_tube_vm,
            )?;
            // early reservation for pass-through PCI devices.
            let endpoint_addr = vfio_pci_device
                .allocate_address(resources)
                .context("failed to allocate resources early for vfio pci dev")?;

            let viommu_mapper = match iommu_dev {
                IommuDevType::NoIommu => None,
                IommuDevType::VirtioIommu => {
                    Some(VfioWrapper::new(vfio_container, vm.get_memory().clone()))
                }
                IommuDevType::CoIommu => {
                    if let Some(endpoints) = coiommu_endpoints {
                        endpoints.push(endpoint_addr.to_u32() as u16);
                    } else {
                        bail!("Missed coiommu_endpoints vector to store the endpoint addr");
                    }
                    None
                }
            };

            if hotplug {
                Ok((VfioDeviceVariant::Pci(vfio_pci_device), None, viommu_mapper))
            } else {
                Ok((
                    VfioDeviceVariant::Pci(vfio_pci_device),
                    simple_jail(jail_config, "vfio_device")?,
                    viommu_mapper,
                ))
            }
        }
        VfioDeviceType::Platform => {
            if guest_address.is_some() {
                bail!("guest-address is not supported for VFIO platform devices");
            }

            if hotplug {
                bail!("hotplug is not supported for VFIO platform devices");
            }

            let vfio_plat_dev =
                VfioPlatformDevice::new(vfio_device, VmMemoryClient::new(vfio_device_tube_mem));

            Ok((
                VfioDeviceVariant::Platform(vfio_plat_dev),
                simple_jail(jail_config, "vfio_platform_device")?,
                None,
            ))
        }
    }
}

/// Setup for devices with virtio-iommu
pub fn setup_virtio_access_platform(
    resources: &mut SystemAllocator,
    iommu_attached_endpoints: &mut BTreeMap<u32, Arc<Mutex<Box<dyn MemoryMapperTrait>>>>,
    devices: &mut [(Box<dyn BusDeviceObj>, Option<Minijail>)],
) -> DeviceResult<(Option<BTreeMap<u32, Tube>>, Option<Tube>)> {
    let mut translate_response_senders: Option<
        BTreeMap<
            u32, // endpoint id
            Tube,
        >,
    > = None;
    let mut tube_pair: Option<(Tube, Tube)> = None;

    for dev in devices.iter_mut() {
        if let Some(pci_dev) = dev.0.as_pci_device_mut() {
            if pci_dev.supports_iommu() {
                let endpoint_id = pci_dev
                    .allocate_address(resources)
                    .context("failed to allocate resources for pci dev")?
                    .to_u32();
                let mapper: Arc<Mutex<Box<dyn MemoryMapperTrait>>> =
                    Arc::new(Mutex::new(Box::new(BasicMemoryMapper::new(u64::MAX))));
                let (request_tx, _request_rx) =
                    tube_pair.get_or_insert_with(|| Tube::pair().unwrap());
                let CreateIpcMapperRet {
                    mapper: ipc_mapper,
                    response_tx,
                } = create_ipc_mapper(
                    endpoint_id,
                    #[allow(deprecated)]
                    request_tx.try_clone()?,
                );
                translate_response_senders
                    .get_or_insert_with(BTreeMap::new)
                    .insert(endpoint_id, response_tx);
                iommu_attached_endpoints.insert(endpoint_id, mapper);
                pci_dev.set_iommu(ipc_mapper)?;
            }
        }
    }

    Ok((
        translate_response_senders,
        tube_pair.map(|(_request_tx, request_rx)| request_rx),
    ))
}