summaryrefslogtreecommitdiffhomepage
path: root/pkg/sentry/fsimpl/proc/task_files.go
blob: 88ea6a6d8134f4eb5fef0fcd564fcea35afa5c3e (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
// Copyright 2019 The gVisor 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
//
//     http://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.

package proc

import (
	"bytes"
	"fmt"
	"io"

	"gvisor.dev/gvisor/pkg/abi/linux"
	"gvisor.dev/gvisor/pkg/context"
	"gvisor.dev/gvisor/pkg/safemem"
	"gvisor.dev/gvisor/pkg/sentry/fsbridge"
	"gvisor.dev/gvisor/pkg/sentry/fsimpl/kernfs"
	"gvisor.dev/gvisor/pkg/sentry/kernel"
	"gvisor.dev/gvisor/pkg/sentry/kernel/auth"
	"gvisor.dev/gvisor/pkg/sentry/limits"
	"gvisor.dev/gvisor/pkg/sentry/mm"
	"gvisor.dev/gvisor/pkg/sentry/usage"
	"gvisor.dev/gvisor/pkg/sentry/vfs"
	"gvisor.dev/gvisor/pkg/syserror"
	"gvisor.dev/gvisor/pkg/usermem"
)

// mm gets the kernel task's MemoryManager. No additional reference is taken on
// mm here. This is safe because MemoryManager.destroy is required to leave the
// MemoryManager in a state where it's still usable as a DynamicBytesSource.
func getMM(task *kernel.Task) *mm.MemoryManager {
	var tmm *mm.MemoryManager
	task.WithMuLocked(func(t *kernel.Task) {
		if mm := t.MemoryManager(); mm != nil {
			tmm = mm
		}
	})
	return tmm
}

// getMMIncRef returns t's MemoryManager. If getMMIncRef succeeds, the
// MemoryManager's users count is incremented, and must be decremented by the
// caller when it is no longer in use.
func getMMIncRef(task *kernel.Task) (*mm.MemoryManager, error) {
	if task.ExitState() == kernel.TaskExitDead {
		return nil, syserror.ESRCH
	}
	var m *mm.MemoryManager
	task.WithMuLocked(func(t *kernel.Task) {
		m = t.MemoryManager()
	})
	if m == nil || !m.IncUsers() {
		return nil, io.EOF
	}
	return m, nil
}

type bufferWriter struct {
	buf *bytes.Buffer
}

// WriteFromBlocks writes up to srcs.NumBytes() bytes from srcs and returns
// the number of bytes written. It may return a partial write without an
// error (i.e. (n, nil) where 0 < n < srcs.NumBytes()). It should not
// return a full write with an error (i.e. srcs.NumBytes(), err) where err
// != nil).
func (w *bufferWriter) WriteFromBlocks(srcs safemem.BlockSeq) (uint64, error) {
	written := srcs.NumBytes()
	for !srcs.IsEmpty() {
		w.buf.Write(srcs.Head().ToSlice())
		srcs = srcs.Tail()
	}
	return written, nil
}

// auxvData implements vfs.DynamicBytesSource for /proc/[pid]/auxv.
//
// +stateify savable
type auxvData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*auxvData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (d *auxvData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	m, err := getMMIncRef(d.task)
	if err != nil {
		return err
	}
	defer m.DecUsers(ctx)

	// Space for buffer with AT_NULL (0) terminator at the end.
	auxv := m.Auxv()
	buf.Grow((len(auxv) + 1) * 16)
	for _, e := range auxv {
		var tmp [8]byte
		usermem.ByteOrder.PutUint64(tmp[:], e.Key)
		buf.Write(tmp[:])

		usermem.ByteOrder.PutUint64(tmp[:], uint64(e.Value))
		buf.Write(tmp[:])
	}
	return nil
}

// execArgType enumerates the types of exec arguments that are exposed through
// proc.
type execArgType int

const (
	cmdlineDataArg execArgType = iota
	environDataArg
)

// cmdlineData implements vfs.DynamicBytesSource for /proc/[pid]/cmdline.
//
// +stateify savable
type cmdlineData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task

	// arg is the type of exec argument this file contains.
	arg execArgType
}

var _ dynamicInode = (*cmdlineData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (d *cmdlineData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	m, err := getMMIncRef(d.task)
	if err != nil {
		return err
	}
	defer m.DecUsers(ctx)

	// Figure out the bounds of the exec arg we are trying to read.
	var ar usermem.AddrRange
	switch d.arg {
	case cmdlineDataArg:
		ar = usermem.AddrRange{
			Start: m.ArgvStart(),
			End:   m.ArgvEnd(),
		}
	case environDataArg:
		ar = usermem.AddrRange{
			Start: m.EnvvStart(),
			End:   m.EnvvEnd(),
		}
	default:
		panic(fmt.Sprintf("unknown exec arg type %v", d.arg))
	}
	if ar.Start == 0 || ar.End == 0 {
		// Don't attempt to read before the start/end are set up.
		return io.EOF
	}

	// N.B. Technically this should be usermem.IOOpts.IgnorePermissions = true
	// until Linux 4.9 (272ddc8b3735 "proc: don't use FOLL_FORCE for reading
	// cmdline and environment").
	writer := &bufferWriter{buf: buf}
	if n, err := m.CopyInTo(ctx, usermem.AddrRangeSeqOf(ar), writer, usermem.IOOpts{}); n == 0 || err != nil {
		// Nothing to copy or something went wrong.
		return err
	}

	// On Linux, if the NULL byte at the end of the argument vector has been
	// overwritten, it continues reading the environment vector as part of
	// the argument vector.
	if d.arg == cmdlineDataArg && buf.Bytes()[buf.Len()-1] != 0 {
		if end := bytes.IndexByte(buf.Bytes(), 0); end != -1 {
			// If we found a NULL character somewhere else in argv, truncate the
			// return up to the NULL terminator (including it).
			buf.Truncate(end)
			return nil
		}

		// There is no NULL terminator in the string, return into envp.
		arEnvv := usermem.AddrRange{
			Start: m.EnvvStart(),
			End:   m.EnvvEnd(),
		}

		// Upstream limits the returned amount to one page of slop.
		// https://elixir.bootlin.com/linux/v4.20/source/fs/proc/base.c#L208
		// we'll return one page total between argv and envp because of the
		// above page restrictions.
		if buf.Len() >= usermem.PageSize {
			// Returned at least one page already, nothing else to add.
			return nil
		}
		remaining := usermem.PageSize - buf.Len()
		if int(arEnvv.Length()) > remaining {
			end, ok := arEnvv.Start.AddLength(uint64(remaining))
			if !ok {
				return syserror.EFAULT
			}
			arEnvv.End = end
		}
		if _, err := m.CopyInTo(ctx, usermem.AddrRangeSeqOf(arEnvv), writer, usermem.IOOpts{}); err != nil {
			return err
		}

		// Linux will return envp up to and including the first NULL character,
		// so find it.
		if end := bytes.IndexByte(buf.Bytes()[ar.Length():], 0); end != -1 {
			buf.Truncate(end)
		}
	}

	return nil
}

// +stateify savable
type commInode struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

func newComm(task *kernel.Task, ino uint64, perm linux.FileMode) *kernfs.Dentry {
	inode := &commInode{task: task}
	inode.DynamicBytesFile.Init(task.Credentials(), ino, &commData{task: task}, perm)

	d := &kernfs.Dentry{}
	d.Init(inode)
	return d
}

func (i *commInode) CheckPermissions(ctx context.Context, creds *auth.Credentials, ats vfs.AccessTypes) error {
	// This file can always be read or written by members of the same thread
	// group. See fs/proc/base.c:proc_tid_comm_permission.
	//
	// N.B. This check is currently a no-op as we don't yet support writing and
	// this file is world-readable anyways.
	t := kernel.TaskFromContext(ctx)
	if t != nil && t.ThreadGroup() == i.task.ThreadGroup() && !ats.MayExec() {
		return nil
	}

	return i.DynamicBytesFile.CheckPermissions(ctx, creds, ats)
}

// commData implements vfs.DynamicBytesSource for /proc/[pid]/comm.
//
// +stateify savable
type commData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*commData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (d *commData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	buf.WriteString(d.task.Name())
	buf.WriteString("\n")
	return nil
}

// idMapData implements vfs.DynamicBytesSource for /proc/[pid]/{gid_map|uid_map}.
//
// +stateify savable
type idMapData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
	gids bool
}

var _ dynamicInode = (*idMapData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (d *idMapData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	var entries []auth.IDMapEntry
	if d.gids {
		entries = d.task.UserNamespace().GIDMap()
	} else {
		entries = d.task.UserNamespace().UIDMap()
	}
	for _, e := range entries {
		fmt.Fprintf(buf, "%10d %10d %10d\n", e.FirstID, e.FirstParentID, e.Length)
	}
	return nil
}

// mapsData implements vfs.DynamicBytesSource for /proc/[pid]/maps.
//
// +stateify savable
type mapsData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*mapsData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (d *mapsData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	if mm := getMM(d.task); mm != nil {
		mm.ReadMapsDataInto(ctx, buf)
	}
	return nil
}

// smapsData implements vfs.DynamicBytesSource for /proc/[pid]/smaps.
//
// +stateify savable
type smapsData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*smapsData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (d *smapsData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	if mm := getMM(d.task); mm != nil {
		mm.ReadSmapsDataInto(ctx, buf)
	}
	return nil
}

// +stateify savable
type taskStatData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task

	// If tgstats is true, accumulate fault stats (not implemented) and CPU
	// time across all tasks in t's thread group.
	tgstats bool

	// pidns is the PID namespace associated with the proc filesystem that
	// includes the file using this statData.
	pidns *kernel.PIDNamespace
}

var _ dynamicInode = (*taskStatData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (s *taskStatData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	fmt.Fprintf(buf, "%d ", s.pidns.IDOfTask(s.task))
	fmt.Fprintf(buf, "(%s) ", s.task.Name())
	fmt.Fprintf(buf, "%c ", s.task.StateStatus()[0])
	ppid := kernel.ThreadID(0)
	if parent := s.task.Parent(); parent != nil {
		ppid = s.pidns.IDOfThreadGroup(parent.ThreadGroup())
	}
	fmt.Fprintf(buf, "%d ", ppid)
	fmt.Fprintf(buf, "%d ", s.pidns.IDOfProcessGroup(s.task.ThreadGroup().ProcessGroup()))
	fmt.Fprintf(buf, "%d ", s.pidns.IDOfSession(s.task.ThreadGroup().Session()))
	fmt.Fprintf(buf, "0 0 " /* tty_nr tpgid */)
	fmt.Fprintf(buf, "0 " /* flags */)
	fmt.Fprintf(buf, "0 0 0 0 " /* minflt cminflt majflt cmajflt */)
	var cputime usage.CPUStats
	if s.tgstats {
		cputime = s.task.ThreadGroup().CPUStats()
	} else {
		cputime = s.task.CPUStats()
	}
	fmt.Fprintf(buf, "%d %d ", linux.ClockTFromDuration(cputime.UserTime), linux.ClockTFromDuration(cputime.SysTime))
	cputime = s.task.ThreadGroup().JoinedChildCPUStats()
	fmt.Fprintf(buf, "%d %d ", linux.ClockTFromDuration(cputime.UserTime), linux.ClockTFromDuration(cputime.SysTime))
	fmt.Fprintf(buf, "%d %d ", s.task.Priority(), s.task.Niceness())
	fmt.Fprintf(buf, "%d ", s.task.ThreadGroup().Count())

	// itrealvalue. Since kernel 2.6.17, this field is no longer
	// maintained, and is hard coded as 0.
	fmt.Fprintf(buf, "0 ")

	// Start time is relative to boot time, expressed in clock ticks.
	fmt.Fprintf(buf, "%d ", linux.ClockTFromDuration(s.task.StartTime().Sub(s.task.Kernel().Timekeeper().BootTime())))

	var vss, rss uint64
	s.task.WithMuLocked(func(t *kernel.Task) {
		if mm := t.MemoryManager(); mm != nil {
			vss = mm.VirtualMemorySize()
			rss = mm.ResidentSetSize()
		}
	})
	fmt.Fprintf(buf, "%d %d ", vss, rss/usermem.PageSize)

	// rsslim.
	fmt.Fprintf(buf, "%d ", s.task.ThreadGroup().Limits().Get(limits.Rss).Cur)

	fmt.Fprintf(buf, "0 0 0 0 0 " /* startcode endcode startstack kstkesp kstkeip */)
	fmt.Fprintf(buf, "0 0 0 0 0 " /* signal blocked sigignore sigcatch wchan */)
	fmt.Fprintf(buf, "0 0 " /* nswap cnswap */)
	terminationSignal := linux.Signal(0)
	if s.task == s.task.ThreadGroup().Leader() {
		terminationSignal = s.task.ThreadGroup().TerminationSignal()
	}
	fmt.Fprintf(buf, "%d ", terminationSignal)
	fmt.Fprintf(buf, "0 0 0 " /* processor rt_priority policy */)
	fmt.Fprintf(buf, "0 0 0 " /* delayacct_blkio_ticks guest_time cguest_time */)
	fmt.Fprintf(buf, "0 0 0 0 0 0 0 " /* start_data end_data start_brk arg_start arg_end env_start env_end */)
	fmt.Fprintf(buf, "0\n" /* exit_code */)

	return nil
}

// statmData implements vfs.DynamicBytesSource for /proc/[pid]/statm.
//
// +stateify savable
type statmData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*statmData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (s *statmData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	var vss, rss uint64
	s.task.WithMuLocked(func(t *kernel.Task) {
		if mm := t.MemoryManager(); mm != nil {
			vss = mm.VirtualMemorySize()
			rss = mm.ResidentSetSize()
		}
	})

	fmt.Fprintf(buf, "%d %d 0 0 0 0 0\n", vss/usermem.PageSize, rss/usermem.PageSize)
	return nil
}

// statusData implements vfs.DynamicBytesSource for /proc/[pid]/status.
//
// +stateify savable
type statusData struct {
	kernfs.DynamicBytesFile

	task  *kernel.Task
	pidns *kernel.PIDNamespace
}

var _ dynamicInode = (*statusData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (s *statusData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	fmt.Fprintf(buf, "Name:\t%s\n", s.task.Name())
	fmt.Fprintf(buf, "State:\t%s\n", s.task.StateStatus())
	fmt.Fprintf(buf, "Tgid:\t%d\n", s.pidns.IDOfThreadGroup(s.task.ThreadGroup()))
	fmt.Fprintf(buf, "Pid:\t%d\n", s.pidns.IDOfTask(s.task))
	ppid := kernel.ThreadID(0)
	if parent := s.task.Parent(); parent != nil {
		ppid = s.pidns.IDOfThreadGroup(parent.ThreadGroup())
	}
	fmt.Fprintf(buf, "PPid:\t%d\n", ppid)
	tpid := kernel.ThreadID(0)
	if tracer := s.task.Tracer(); tracer != nil {
		tpid = s.pidns.IDOfTask(tracer)
	}
	fmt.Fprintf(buf, "TracerPid:\t%d\n", tpid)
	var fds int
	var vss, rss, data uint64
	s.task.WithMuLocked(func(t *kernel.Task) {
		if fdTable := t.FDTable(); fdTable != nil {
			fds = fdTable.Size()
		}
		if mm := t.MemoryManager(); mm != nil {
			vss = mm.VirtualMemorySize()
			rss = mm.ResidentSetSize()
			data = mm.VirtualDataSize()
		}
	})
	fmt.Fprintf(buf, "FDSize:\t%d\n", fds)
	fmt.Fprintf(buf, "VmSize:\t%d kB\n", vss>>10)
	fmt.Fprintf(buf, "VmRSS:\t%d kB\n", rss>>10)
	fmt.Fprintf(buf, "VmData:\t%d kB\n", data>>10)
	fmt.Fprintf(buf, "Threads:\t%d\n", s.task.ThreadGroup().Count())
	creds := s.task.Credentials()
	fmt.Fprintf(buf, "CapInh:\t%016x\n", creds.InheritableCaps)
	fmt.Fprintf(buf, "CapPrm:\t%016x\n", creds.PermittedCaps)
	fmt.Fprintf(buf, "CapEff:\t%016x\n", creds.EffectiveCaps)
	fmt.Fprintf(buf, "CapBnd:\t%016x\n", creds.BoundingCaps)
	fmt.Fprintf(buf, "Seccomp:\t%d\n", s.task.SeccompMode())
	// We unconditionally report a single NUMA node. See
	// pkg/sentry/syscalls/linux/sys_mempolicy.go.
	fmt.Fprintf(buf, "Mems_allowed:\t1\n")
	fmt.Fprintf(buf, "Mems_allowed_list:\t0\n")
	return nil
}

// ioUsage is the /proc/[pid]/io and /proc/[pid]/task/[tid]/io data provider.
type ioUsage interface {
	// IOUsage returns the io usage data.
	IOUsage() *usage.IO
}

// +stateify savable
type ioData struct {
	kernfs.DynamicBytesFile

	ioUsage
}

var _ dynamicInode = (*ioData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (i *ioData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	io := usage.IO{}
	io.Accumulate(i.IOUsage())

	fmt.Fprintf(buf, "char: %d\n", io.CharsRead)
	fmt.Fprintf(buf, "wchar: %d\n", io.CharsWritten)
	fmt.Fprintf(buf, "syscr: %d\n", io.ReadSyscalls)
	fmt.Fprintf(buf, "syscw: %d\n", io.WriteSyscalls)
	fmt.Fprintf(buf, "read_bytes: %d\n", io.BytesRead)
	fmt.Fprintf(buf, "write_bytes: %d\n", io.BytesWritten)
	fmt.Fprintf(buf, "cancelled_write_bytes: %d\n", io.BytesWriteCancelled)
	return nil
}

// oomScoreAdj is a stub of the /proc/<pid>/oom_score_adj file.
//
// +stateify savable
type oomScoreAdj struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ vfs.WritableDynamicBytesSource = (*oomScoreAdj)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (o *oomScoreAdj) Generate(ctx context.Context, buf *bytes.Buffer) error {
	if o.task.ExitState() == kernel.TaskExitDead {
		return syserror.ESRCH
	}
	fmt.Fprintf(buf, "%d\n", o.task.OOMScoreAdj())
	return nil
}

// Write implements vfs.WritableDynamicBytesSource.Write.
func (o *oomScoreAdj) Write(ctx context.Context, src usermem.IOSequence, offset int64) (int64, error) {
	if src.NumBytes() == 0 {
		return 0, nil
	}

	// Limit input size so as not to impact performance if input size is large.
	src = src.TakeFirst(usermem.PageSize - 1)

	var v int32
	n, err := usermem.CopyInt32StringInVec(ctx, src.IO, src.Addrs, &v, src.Opts)
	if err != nil {
		return 0, err
	}

	if o.task.ExitState() == kernel.TaskExitDead {
		return 0, syserror.ESRCH
	}
	if err := o.task.SetOOMScoreAdj(v); err != nil {
		return 0, err
	}

	return n, nil
}

// exeSymlink is an symlink for the /proc/[pid]/exe file.
//
// +stateify savable
type exeSymlink struct {
	kernfs.InodeAttrs
	kernfs.InodeNoopRefCount
	kernfs.InodeSymlink

	task *kernel.Task
}

var _ kernfs.Inode = (*exeSymlink)(nil)

func newExeSymlink(task *kernel.Task, ino uint64) *kernfs.Dentry {
	inode := &exeSymlink{task: task}
	inode.Init(task.Credentials(), ino, linux.ModeSymlink|0777)

	d := &kernfs.Dentry{}
	d.Init(inode)
	return d
}

// Readlink implements kernfs.Inode.
func (s *exeSymlink) Readlink(ctx context.Context) (string, error) {
	if !kernel.ContextCanTrace(ctx, s.task, false) {
		return "", syserror.EACCES
	}

	// Pull out the executable for /proc/[pid]/exe.
	exec, err := s.executable()
	if err != nil {
		return "", err
	}
	defer exec.DecRef()

	return exec.PathnameWithDeleted(ctx), nil
}

// Getlink implements kernfs.Inode.Getlink.
func (s *exeSymlink) Getlink(ctx context.Context) (vfs.VirtualDentry, string, error) {
	if !kernel.ContextCanTrace(ctx, s.task, false) {
		return vfs.VirtualDentry{}, "", syserror.EACCES
	}

	exec, err := s.executable()
	if err != nil {
		return vfs.VirtualDentry{}, "", err
	}
	defer exec.DecRef()

	vd := exec.(*fsbridge.VFSFile).FileDescription().VirtualDentry()
	vd.IncRef()
	return vd, "", nil
}

func (s *exeSymlink) executable() (file fsbridge.File, err error) {
	s.task.WithMuLocked(func(t *kernel.Task) {
		mm := t.MemoryManager()
		if mm == nil {
			// TODO(b/34851096): Check shouldn't allow Readlink once the
			// Task is zombied.
			err = syserror.EACCES
			return
		}

		// The MemoryManager may be destroyed, in which case
		// MemoryManager.destroy will simply set the executable to nil
		// (with locks held).
		file = mm.Executable()
		if file == nil {
			err = syserror.ENOENT
		}
	})
	return
}

// mountInfoData is used to implement /proc/[pid]/mountinfo.
//
// +stateify savable
type mountInfoData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*mountInfoData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (i *mountInfoData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	var fsctx *kernel.FSContext
	i.task.WithMuLocked(func(t *kernel.Task) {
		fsctx = t.FSContext()
	})
	if fsctx == nil {
		// The task has been destroyed. Nothing to show here.
		return nil
	}
	rootDir := fsctx.RootDirectoryVFS2()
	if !rootDir.Ok() {
		// Root has been destroyed. Don't try to read mounts.
		return nil
	}
	defer rootDir.DecRef()
	i.task.Kernel().VFS().GenerateProcMountInfo(ctx, rootDir, buf)
	return nil
}

// mountsData is used to implement /proc/[pid]/mounts.
//
// +stateify savable
type mountsData struct {
	kernfs.DynamicBytesFile

	task *kernel.Task
}

var _ dynamicInode = (*mountsData)(nil)

// Generate implements vfs.DynamicBytesSource.Generate.
func (i *mountsData) Generate(ctx context.Context, buf *bytes.Buffer) error {
	var fsctx *kernel.FSContext
	i.task.WithMuLocked(func(t *kernel.Task) {
		fsctx = t.FSContext()
	})
	if fsctx == nil {
		// The task has been destroyed. Nothing to show here.
		return nil
	}
	rootDir := fsctx.RootDirectoryVFS2()
	if !rootDir.Ok() {
		// Root has been destroyed. Don't try to read mounts.
		return nil
	}
	defer rootDir.DecRef()
	i.task.Kernel().VFS().GenerateProcMounts(ctx, rootDir, buf)
	return nil
}