forked from apple/containerization
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathLinuxContainer.swift
More file actions
909 lines (806 loc) · 32.7 KB
/
LinuxContainer.swift
File metadata and controls
909 lines (806 loc) · 32.7 KB
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
//===----------------------------------------------------------------------===//
// Copyright © 2025 Apple Inc. and the Containerization project authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// https://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//===----------------------------------------------------------------------===//
#if os(macOS)
import ContainerizationError
import ContainerizationExtras
import ContainerizationOCI
import Foundation
import Logging
import Synchronization
import struct ContainerizationOS.Terminal
/// `LinuxContainer` is an easy to use type for launching and managing the
/// full lifecycle of a Linux container ran inside of a virtual machine.
public final class LinuxContainer: Container, Sendable {
/// The identifier of the container.
public let id: String
/// Rootfs for the container.
public let rootfs: Mount
/// Configuration for the container.
public let config: Configuration
/// The configuration for the LinuxContainer.
public struct Configuration: Sendable {
/// Configuration for the init process of the container.
public var process = LinuxProcessConfiguration()
/// The amount of cpus for the container.
public var cpus: Int = 4
/// The memory in bytes to give to the container.
public var memoryInBytes: UInt64 = 1024.mib()
/// Requested storage in bytes.
public var storageInBytes: UInt64?
/// The hostname for the container.
public var hostname: String = ""
/// The system control options for the container.
public var sysctl: [String: String] = [:]
/// The network interfaces for the container.
public var interfaces: [any Interface] = []
/// The Unix domain socket relays to setup for the container.
public var sockets: [UnixSocketConfiguration] = []
/// The mounts for the container.
public var mounts: [Mount] = LinuxContainer.defaultMounts()
/// The DNS configuration for the container.
public var dns: DNS?
/// The hosts to add to /etc/hosts for the container.
public var hosts: Hosts?
/// Enable nested virtualization support.
public var virtualization: Bool = false
/// Optional destination for serial boot logs.
public var bootLog: BootLog?
/// EXPERIMENTAL: Path in the root filesystem for the virtual
/// machine where the OCI runtime used to spawn the container lives.
public var ociRuntimePath: String?
public init() {}
public init(
process: LinuxProcessConfiguration,
cpus: Int = 4,
memoryInBytes: UInt64 = 1024.mib(),
storageInBytes: UInt64? = nil,
hostname: String = "",
sysctl: [String: String] = [:],
interfaces: [any Interface] = [],
sockets: [UnixSocketConfiguration] = [],
mounts: [Mount] = LinuxContainer.defaultMounts(),
dns: DNS? = nil,
hosts: Hosts? = nil,
virtualization: Bool = false,
bootLog: BootLog? = nil,
ociRuntimePath: String? = nil
) {
self.process = process
self.cpus = cpus
self.memoryInBytes = memoryInBytes
self.storageInBytes = storageInBytes
self.hostname = hostname
self.sysctl = sysctl
self.interfaces = interfaces
self.sockets = sockets
self.mounts = mounts
self.dns = dns
self.hosts = hosts
self.virtualization = virtualization
self.bootLog = bootLog
self.ociRuntimePath = ociRuntimePath
}
}
private let state: AsyncMutex<State>
// Ports to be allocated from for stdio and for
// unix socket relays that are sharing a guest
// uds to the host.
private let hostVsockPorts: Atomic<UInt32>
// Ports we request the guest to allocate for unix socket relays from
// the host.
private let guestVsockPorts: Atomic<UInt32>
private enum State: Sendable {
/// The container class has been created but no live resources are running.
case initialized
/// The container's virtual machine has been setup and the runtime environment has been configured.
case created(CreatedState)
/// The initial process of the container has started and is running.
case started(StartedState)
/// The container has run and fully stopped.
case stopped
/// An error occurred during the lifetime of this class.
case errored(Swift.Error)
/// The container is paused.
case paused(PausedState)
struct CreatedState: Sendable {
let vm: any VirtualMachineInstance
let relayManager: UnixSocketRelayManager
}
struct StartedState: Sendable {
let vm: any VirtualMachineInstance
let process: LinuxProcess
let relayManager: UnixSocketRelayManager
var vendedProcesses: [String: LinuxProcess]
init(_ state: CreatedState, process: LinuxProcess) {
self.vm = state.vm
self.relayManager = state.relayManager
self.process = process
self.vendedProcesses = [:]
}
init(_ state: PausedState) {
self.vm = state.vm
self.relayManager = state.relayManager
self.process = state.process
self.vendedProcesses = state.vendedProcesses
}
}
struct PausedState: Sendable {
let vm: any VirtualMachineInstance
let relayManager: UnixSocketRelayManager
let process: LinuxProcess
var vendedProcesses: [String: LinuxProcess]
init(_ state: StartedState) {
self.vm = state.vm
self.relayManager = state.relayManager
self.process = state.process
self.vendedProcesses = state.vendedProcesses
}
}
func createdState(_ operation: String) throws -> CreatedState {
switch self {
case .created(let state):
return state
case .errored(let err):
throw err
default:
throw ContainerizationError(
.invalidState,
message: "failed to \(operation): container must be created"
)
}
}
func startedState(_ operation: String) throws -> StartedState {
switch self {
case .started(let state):
return state
case .errored(let err):
throw err
default:
throw ContainerizationError(
.invalidState,
message: "failed to \(operation): container must be running"
)
}
}
func pausedState(_ operation: String) throws -> PausedState {
switch self {
case .paused(let state):
return state
case .errored(let err):
throw err
default:
throw ContainerizationError(
.invalidState,
message: "failed to \(operation): container must be paused"
)
}
}
mutating func validateForCreate() throws {
switch self {
case .initialized, .stopped:
break
case .errored(let err):
throw err
default:
throw ContainerizationError(
.invalidState,
message: "container must be in initialized or stopped state to create"
)
}
}
mutating func setErrored(error: Swift.Error) {
self = .errored(error)
}
}
private let vmm: VirtualMachineManager
private let logger: Logger?
/// Create a new `LinuxContainer`.
///
/// - Parameters:
/// - id: The identifier for the container.
/// - rootfs: The root filesystem mount containing the container image contents.
/// - vmm: The virtual machine manager that will handle launching the VM for the container.
/// - logger: Optional logger for container operations.
/// - configuration: A closure that configures the container by modifying the Configuration instance.
public init(
_ id: String,
rootfs: Mount,
vmm: VirtualMachineManager,
logger: Logger? = nil,
configuration: (inout Configuration) throws -> Void
) throws {
self.id = id
self.vmm = vmm
self.hostVsockPorts = Atomic<UInt32>(0x1000_0000)
self.guestVsockPorts = Atomic<UInt32>(0x1000_0000)
self.rootfs = rootfs
self.logger = logger
var config = Configuration()
try configuration(&config)
self.config = config
self.state = AsyncMutex(.initialized)
}
/// Create a new `LinuxContainer`.
///
/// - Parameters:
/// - id: The identifier for the container.
/// - rootfs: The root filesystem mount containing the container image contents.
/// - vmm: The virtual machine manager that will handle launching the VM for the container.
/// - configuration: The container configuration specifying process, resources, networking, and other settings.
/// - logger: Optional logger for container operations.
public init(
_ id: String,
rootfs: Mount,
vmm: VirtualMachineManager,
configuration: LinuxContainer.Configuration,
logger: Logger? = nil
) {
self.id = id
self.vmm = vmm
self.hostVsockPorts = Atomic<UInt32>(0x1000_0000)
self.guestVsockPorts = Atomic<UInt32>(0x1000_0000)
self.rootfs = rootfs
self.logger = logger
self.config = configuration
self.state = AsyncMutex(.initialized)
}
private static func createDefaultRuntimeSpec(_ id: String) -> Spec {
.init(
process: .init(),
hostname: id,
root: .init(
path: Self.guestRootfsPath(id),
readonly: false
),
linux: .init(
resources: .init(),
cgroupsPath: "/container/\(id)"
)
)
}
private func generateRuntimeSpec() -> Spec {
var spec = Self.createDefaultRuntimeSpec(id)
// Process toggles.
spec.process = config.process.toOCI()
// General toggles.
spec.hostname = config.hostname
// Linux toggles.
spec.linux?.sysctl = config.sysctl
// Resource limits.
// CPU: quota/period model where period is 100ms (100,000µs) and quota is cpus * period
// Memory: limit in bytes
spec.linux?.resources = LinuxResources(
memory: LinuxMemory(
limit: Int64(config.memoryInBytes)
),
cpu: LinuxCPU(
quota: Int64(config.cpus * 100_000),
period: 100_000
)
)
spec.linux?.namespaces = [
LinuxNamespace(type: .cgroup),
LinuxNamespace(type: .ipc),
LinuxNamespace(type: .mount),
LinuxNamespace(type: .pid),
LinuxNamespace(type: .uts),
]
return spec
}
/// The default set of mounts for a LinuxContainer.
public static func defaultMounts() -> [Mount] {
let defaultOptions = ["nosuid", "noexec", "nodev"]
return [
.any(type: "proc", source: "proc", destination: "/proc"),
.any(type: "sysfs", source: "sysfs", destination: "/sys", options: defaultOptions),
.any(type: "devtmpfs", source: "none", destination: "/dev", options: ["nosuid", "mode=755"]),
.any(type: "mqueue", source: "mqueue", destination: "/dev/mqueue", options: defaultOptions),
.any(type: "tmpfs", source: "tmpfs", destination: "/dev/shm", options: defaultOptions + ["mode=1777", "size=65536k"]),
.any(type: "cgroup2", source: "none", destination: "/sys/fs/cgroup", options: defaultOptions),
.any(type: "devpts", source: "devpts", destination: "/dev/pts", options: ["nosuid", "noexec", "newinstance", "gid=5", "mode=0620", "ptmxmode=0666"]),
]
}
/// A more traditional default set of mounts that OCI runtimes typically employ.
public static func defaultOCIMounts() -> [Mount] {
let defaultOptions = ["nosuid", "noexec", "nodev"]
return [
.any(type: "proc", source: "proc", destination: "/proc"),
.any(type: "tmpfs", source: "tmpfs", destination: "/dev", options: ["nosuid", "mode=755", "size=65536k"]),
.any(type: "devpts", source: "devpts", destination: "/dev/pts", options: ["nosuid", "noexec", "newinstance", "gid=5", "mode=0620", "ptmxmode=0666"]),
.any(type: "sysfs", source: "sysfs", destination: "/sys", options: defaultOptions),
.any(type: "mqueue", source: "mqueue", destination: "/dev/mqueue", options: defaultOptions),
.any(type: "tmpfs", source: "tmpfs", destination: "/dev/shm", options: defaultOptions + ["mode=1777", "size=65536k"]),
.any(type: "cgroup2", source: "none", destination: "/sys/fs/cgroup", options: defaultOptions),
]
}
private static func guestRootfsPath(_ id: String) -> String {
"/run/container/\(id)/rootfs"
}
}
extension LinuxContainer {
package var root: String {
Self.guestRootfsPath(id)
}
/// Number of CPU cores allocated.
public var cpus: Int {
config.cpus
}
/// Amount of memory in bytes allocated for the container.
/// This will be aligned to a 1MB boundary if it isn't already.
public var memoryInBytes: UInt64 {
config.memoryInBytes
}
/// Network interfaces of the container.
public var interfaces: [any Interface] {
config.interfaces
}
/// Create and start the underlying container's virtual machine
/// and set up the runtime environment. The container's init process
/// is NOT running afterwards.
public func create() async throws {
try await self.state.withLock { state in
try state.validateForCreate()
let vmConfig = VMConfiguration(
cpus: self.cpus,
memoryInBytes: self.memoryInBytes,
diskStorageBytes: self.config.storageInBytes,
interfaces: self.interfaces,
mountsByID: [self.id: [self.rootfs] + self.config.mounts],
bootLog: self.config.bootLog,
nestedVirtualization: self.config.virtualization
)
let creationConfig = StandardVMConfig(configuration: vmConfig)
let vm = try await self.vmm.create(config: creationConfig)
let relayManager = UnixSocketRelayManager(vm: vm, log: self.logger)
try await vm.start()
do {
try await vm.withAgent { agent in
try await agent.standardSetup()
// Mount the rootfs.
guard let attachments = vm.mounts[self.id], let rootfsAttachment = attachments.first else {
throw ContainerizationError(.notFound, message: "rootfs mount not found")
}
var rootfs = rootfsAttachment.to
rootfs.destination = Self.guestRootfsPath(self.id)
try await agent.mount(rootfs)
// Start up our friendly unix socket relays.
for socket in self.config.sockets {
try await self.relayUnixSocket(
socket: socket,
relayManager: relayManager,
agent: agent
)
}
// For every interface asked for:
// 1. Add the address requested
// 2. Online the adapter
// 3. If a gateway IP address is present, add the default route.
for (index, i) in self.interfaces.enumerated() {
let name = "eth\(index)"
try await agent.addressAdd(name: name, address: i.address)
try await agent.up(name: name, mtu: 1280)
if let gateway = i.gateway {
try await agent.routeAddDefault(name: name, gateway: gateway)
}
}
// Setup /etc/resolv.conf and /etc/hosts if asked for.
if let dns = self.config.dns {
try await agent.configureDNS(config: dns, location: rootfs.destination)
}
if let hosts = self.config.hosts {
try await agent.configureHosts(config: hosts, location: rootfs.destination)
}
}
state = .created(.init(vm: vm, relayManager: relayManager))
} catch {
try? await relayManager.stopAll()
try? await vm.stop()
state.setErrored(error: error)
throw error
}
}
}
/// Start the container's initial process.
public func start() async throws {
try await self.state.withLock { state in
let createdState = try state.createdState("start")
let agent = try await createdState.vm.dialAgent()
do {
var spec = self.generateRuntimeSpec()
// We don't need the rootfs, nor do OCI runtimes want it included.
let containerMounts = createdState.vm.mounts[self.id] ?? []
spec.mounts = containerMounts.dropFirst().map { $0.to }
let stdio = IOUtil.setup(
portAllocator: self.hostVsockPorts,
stdin: self.config.process.stdin,
stdout: self.config.process.stdout,
stderr: self.config.process.stderr
)
let process = LinuxProcess(
self.id,
containerID: self.id,
spec: spec,
io: stdio,
ociRuntimePath: self.config.ociRuntimePath,
agent: agent,
vm: createdState.vm,
logger: self.logger
)
try await process.start()
state = .started(.init(createdState, process: process))
} catch {
try? await agent.close()
try? await createdState.vm.stop()
state.setErrored(error: error)
throw error
}
}
}
private static func setupIO(
portAllocator: borrowing Atomic<UInt32>,
stdin: ReaderStream?,
stdout: Writer?,
stderr: Writer?
) -> LinuxProcess.Stdio {
var stdinSetup: LinuxProcess.StdioReaderSetup? = nil
if let reader = stdin {
let ret = portAllocator.wrappingAdd(1, ordering: .relaxed)
stdinSetup = .init(
port: ret.oldValue,
reader: reader
)
}
var stdoutSetup: LinuxProcess.StdioSetup? = nil
if let writer = stdout {
let ret = portAllocator.wrappingAdd(1, ordering: .relaxed)
stdoutSetup = LinuxProcess.StdioSetup(
port: ret.oldValue,
writer: writer
)
}
var stderrSetup: LinuxProcess.StdioSetup? = nil
if let writer = stderr {
let ret = portAllocator.wrappingAdd(1, ordering: .relaxed)
stderrSetup = LinuxProcess.StdioSetup(
port: ret.oldValue,
writer: writer
)
}
return LinuxProcess.Stdio(
stdin: stdinSetup,
stdout: stdoutSetup,
stderr: stderrSetup
)
}
/// Stop the container from executing. This MUST be called even if wait() has returned
/// as their are additional resources to free.
public func stop() async throws {
try await self.state.withLock { state in
// Allow stop to be called multiple times.
if case .stopped = state {
return
}
let startedState = try state.startedState("stop")
let vm = startedState.vm
var firstError: Error?
do {
try await startedState.relayManager.stopAll()
} catch {
self.logger?.error("failed to stop relay manager: \(error)")
firstError = firstError ?? error
}
do {
try await vm.withAgent { agent in
// First, we need to stop any unix socket relays as this will
// keep the rootfs from being able to umount (EBUSY).
let sockets = self.config.sockets
if !sockets.isEmpty {
guard let relayAgent = agent as? SocketRelayAgent else {
throw ContainerizationError(
.unsupported,
message: "VirtualMachineAgent does not support relaySocket surface"
)
}
for socket in sockets {
try await relayAgent.stopSocketRelay(configuration: socket)
}
}
// Now lets ensure every process is donezo.
try await agent.kill(pid: -1, signal: SIGKILL)
// Wait on init proc exit. Give it 5 seconds of leeway.
_ = try await agent.waitProcess(
id: self.id,
containerID: self.id,
timeoutInSeconds: 5
)
// Today, we leave EBUSY looping and other fun logic up to the
// guest agent.
try await agent.umount(
path: Self.guestRootfsPath(self.id),
flags: 0
)
try await agent.sync()
}
} catch {
self.logger?.error("failed during guest cleanup: \(error)")
firstError = firstError ?? error
}
for process in startedState.vendedProcesses.values {
do {
try await process._delete()
} catch {
self.logger?.error("failed to delete process \(process.id): \(error)")
firstError = firstError ?? error
}
}
do {
try await startedState.process.delete()
} catch {
self.logger?.error("failed to delete init process: \(error)")
firstError = firstError ?? error
}
do {
try await vm.stop()
state = .stopped
if let firstError {
throw firstError
}
} catch {
self.logger?.error("failed to stop VM: \(error)")
let finalError = firstError ?? error
state.setErrored(error: finalError)
throw finalError
}
}
}
/// Send a signal to the container.
public func kill(_ signal: Int32) async throws {
try await self.state.withLock {
let state = try $0.startedState("kill")
try await state.process.kill(signal)
}
}
/// Wait for the container to exit. Returns the exit code.
@discardableResult
public func wait(timeoutInSeconds: Int64? = nil) async throws -> ExitStatus {
let t = try await self.state.withLock {
let state = try $0.startedState("wait")
let t = Task {
try await state.process.wait(timeoutInSeconds: timeoutInSeconds)
}
return t
}
return try await t.value
}
/// Resize the container's terminal (if one was requested). This
/// will error if terminal was set to false before creating the container.
public func resize(to: Terminal.Size) async throws {
try await self.state.withLock {
let state = try $0.startedState("resize")
try await state.process.resize(to: to)
}
}
/// Execute a new process in the container. The process is not started after this call, and must be manually started
/// via the `start` method.
public func exec(_ id: String, configuration: @Sendable @escaping (inout LinuxProcessConfiguration) throws -> Void) async throws -> LinuxProcess {
try await self.state.withLock { state in
var startedState = try state.startedState("exec")
var spec = self.generateRuntimeSpec()
var config = LinuxProcessConfiguration()
try configuration(&config)
spec.process = config.toOCI()
let stdio = IOUtil.setup(
portAllocator: self.hostVsockPorts,
stdin: config.stdin,
stdout: config.stdout,
stderr: config.stderr
)
let agent = try await startedState.vm.dialAgent()
let process = LinuxProcess(
id,
containerID: self.id,
spec: spec,
io: stdio,
ociRuntimePath: self.config.ociRuntimePath,
agent: agent,
vm: startedState.vm,
logger: self.logger,
onDelete: { [weak self] in
await self?.removeProcess(id: id)
}
)
startedState.vendedProcesses[id] = process
state = .started(startedState)
return process
}
}
/// Execute a new process in the container. The process is not started after this call, and must be manually started
/// via the `start` method.
public func exec(_ id: String, configuration: LinuxProcessConfiguration) async throws -> LinuxProcess {
try await self.state.withLock {
var state = try $0.startedState("exec")
var spec = self.generateRuntimeSpec()
spec.process = configuration.toOCI()
let stdio = IOUtil.setup(
portAllocator: self.hostVsockPorts,
stdin: configuration.stdin,
stdout: configuration.stdout,
stderr: configuration.stderr
)
let agent = try await state.vm.dialAgent()
let process = LinuxProcess(
id,
containerID: self.id,
spec: spec,
io: stdio,
ociRuntimePath: self.config.ociRuntimePath,
agent: agent,
vm: state.vm,
logger: self.logger,
onDelete: { [weak self] in
await self?.removeProcess(id: id)
}
)
state.vendedProcesses[id] = process
$0 = .started(state)
return process
}
}
/// Dial a vsock port in the container.
public func dialVsock(port: UInt32) async throws -> FileHandle {
try await self.state.withLock {
let state = try $0.startedState("dialVsock")
return try await state.vm.dial(port)
}
}
/// Close the containers standard input to signal no more input is
/// arriving.
public func closeStdin() async throws {
try await self.state.withLock {
let state = try $0.startedState("closeStdin")
return try await state.process.closeStdin()
}
}
/// Remove a process from the vended processes tracking.
private func removeProcess(id: String) async {
await self.state.withLock {
guard case .started(var state) = $0 else {
return
}
state.vendedProcesses.removeValue(forKey: id)
$0 = .started(state)
}
}
/// Get statistics for the container.
public func statistics() async throws -> ContainerStatistics {
try await self.state.withLock {
let state = try $0.startedState("statistics")
let stats = try await state.vm.withAgent { agent in
let allStats = try await agent.containerStatistics(containerIDs: [self.id])
guard let containerStats = allStats.first else {
throw ContainerizationError(
.notFound,
message: "statistics for container \(self.id) not found"
)
}
return containerStats
}
return stats
}
}
/// Relay a unix socket from in the container to the host, or from the host
/// to inside the container.
public func relayUnixSocket(socket: UnixSocketConfiguration) async throws {
try await self.state.withLock {
let state = try $0.startedState("relayUnixSocket")
try await state.vm.withAgent { agent in
try await self.relayUnixSocket(
socket: socket,
relayManager: state.relayManager,
agent: agent
)
}
}
}
private func relayUnixSocket(
socket: UnixSocketConfiguration,
relayManager: UnixSocketRelayManager,
agent: any VirtualMachineAgent
) async throws {
guard let relayAgent = agent as? SocketRelayAgent else {
throw ContainerizationError(
.unsupported,
message: "VirtualMachineAgent does not support relaySocket surface"
)
}
var socket = socket
let rootInGuest = URL(filePath: self.root)
let port: UInt32
if socket.direction == .into {
port = self.hostVsockPorts.wrappingAdd(1, ordering: .relaxed).oldValue
socket.destination = rootInGuest.appending(path: socket.destination.path)
} else {
port = self.guestVsockPorts.wrappingAdd(1, ordering: .relaxed).oldValue
socket.source = rootInGuest.appending(path: socket.source.path)
}
try await relayManager.start(port: port, socket: socket)
try await relayAgent.relaySocket(port: port, configuration: socket)
}
}
extension VirtualMachineInstance {
/// Scoped access to an agent instance to ensure the resources are always freed (mostly close(2)'ing
/// the vsock fd)
func withAgent<T>(fn: @Sendable (VirtualMachineAgent) async throws -> T) async throws -> T {
let agent = try await self.dialAgent()
do {
let result = try await fn(agent)
try await agent.close()
return result
} catch {
try? await agent.close()
throw error
}
}
}
extension AttachedFilesystem {
var to: ContainerizationOCI.Mount {
.init(
type: self.type,
source: self.source,
destination: self.destination,
options: self.options
)
}
}
struct IOUtil {
static func setup(
portAllocator: borrowing Atomic<UInt32>,
stdin: ReaderStream?,
stdout: Writer?,
stderr: Writer?
) -> LinuxProcess.Stdio {
var stdinSetup: LinuxProcess.StdioReaderSetup? = nil
if let reader = stdin {
let ret = portAllocator.wrappingAdd(1, ordering: .relaxed)
stdinSetup = .init(
port: ret.oldValue,
reader: reader
)
}
var stdoutSetup: LinuxProcess.StdioSetup? = nil
if let writer = stdout {
let ret = portAllocator.wrappingAdd(1, ordering: .relaxed)
stdoutSetup = LinuxProcess.StdioSetup(
port: ret.oldValue,
writer: writer
)
}
var stderrSetup: LinuxProcess.StdioSetup? = nil
if let writer = stderr {
let ret = portAllocator.wrappingAdd(1, ordering: .relaxed)
stderrSetup = LinuxProcess.StdioSetup(
port: ret.oldValue,
writer: writer
)
}
return LinuxProcess.Stdio(
stdin: stdinSetup,
stdout: stdoutSetup,
stderr: stderrSetup
)
}
}
#endif