summaryrefslogtreecommitdiffhomepage
path: root/pkg/sentry/syscalls/linux/vfs2/filesystem.go
blob: a859095e2859baafdeb4e163ee092686e59704b9 (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
// Copyright 2020 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 vfs2

import (
	"gvisor.dev/gvisor/pkg/abi/linux"
	"gvisor.dev/gvisor/pkg/sentry/arch"
	"gvisor.dev/gvisor/pkg/sentry/kernel"
	"gvisor.dev/gvisor/pkg/sentry/vfs"
	"gvisor.dev/gvisor/pkg/syserror"
	"gvisor.dev/gvisor/pkg/usermem"
)

// Link implements Linux syscall link(2).
func Link(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	oldpathAddr := args[0].Pointer()
	newpathAddr := args[1].Pointer()
	return 0, nil, linkat(t, linux.AT_FDCWD, oldpathAddr, linux.AT_FDCWD, newpathAddr, 0 /* flags */)
}

// Linkat implements Linux syscall linkat(2).
func Linkat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	olddirfd := args[0].Int()
	oldpathAddr := args[1].Pointer()
	newdirfd := args[2].Int()
	newpathAddr := args[3].Pointer()
	flags := args[4].Int()
	return 0, nil, linkat(t, olddirfd, oldpathAddr, newdirfd, newpathAddr, flags)
}

func linkat(t *kernel.Task, olddirfd int32, oldpathAddr usermem.Addr, newdirfd int32, newpathAddr usermem.Addr, flags int32) error {
	if flags&^(linux.AT_EMPTY_PATH|linux.AT_SYMLINK_FOLLOW) != 0 {
		return syserror.EINVAL
	}
	if flags&linux.AT_EMPTY_PATH != 0 && !t.HasCapability(linux.CAP_DAC_READ_SEARCH) {
		return syserror.ENOENT
	}

	oldpath, err := copyInPath(t, oldpathAddr)
	if err != nil {
		return err
	}
	oldtpop, err := getTaskPathOperation(t, olddirfd, oldpath, shouldAllowEmptyPath(flags&linux.AT_EMPTY_PATH != 0), shouldFollowFinalSymlink(flags&linux.AT_SYMLINK_FOLLOW != 0))
	if err != nil {
		return err
	}
	defer oldtpop.Release()

	newpath, err := copyInPath(t, newpathAddr)
	if err != nil {
		return err
	}
	newtpop, err := getTaskPathOperation(t, newdirfd, newpath, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer newtpop.Release()

	return t.Kernel().VFS().LinkAt(t, t.Credentials(), &oldtpop.pop, &newtpop.pop)
}

// Mkdir implements Linux syscall mkdir(2).
func Mkdir(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	addr := args[0].Pointer()
	mode := args[1].ModeT()
	return 0, nil, mkdirat(t, linux.AT_FDCWD, addr, mode)
}

// Mkdirat implements Linux syscall mkdirat(2).
func Mkdirat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	dirfd := args[0].Int()
	addr := args[1].Pointer()
	mode := args[2].ModeT()
	return 0, nil, mkdirat(t, dirfd, addr, mode)
}

func mkdirat(t *kernel.Task, dirfd int32, addr usermem.Addr, mode uint) error {
	path, err := copyInPath(t, addr)
	if err != nil {
		return err
	}
	tpop, err := getTaskPathOperation(t, dirfd, path, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer tpop.Release()
	return t.Kernel().VFS().MkdirAt(t, t.Credentials(), &tpop.pop, &vfs.MkdirOptions{
		Mode: linux.FileMode(mode & (0777 | linux.S_ISVTX) &^ t.FSContext().Umask()),
	})
}

// Mknod implements Linux syscall mknod(2).
func Mknod(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	addr := args[0].Pointer()
	mode := args[1].ModeT()
	dev := args[2].Uint()
	return 0, nil, mknodat(t, linux.AT_FDCWD, addr, mode, dev)
}

// Mknodat implements Linux syscall mknodat(2).
func Mknodat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	dirfd := args[0].Int()
	addr := args[1].Pointer()
	mode := args[2].ModeT()
	dev := args[3].Uint()
	return 0, nil, mknodat(t, dirfd, addr, mode, dev)
}

func mknodat(t *kernel.Task, dirfd int32, addr usermem.Addr, mode uint, dev uint32) error {
	path, err := copyInPath(t, addr)
	if err != nil {
		return err
	}
	tpop, err := getTaskPathOperation(t, dirfd, path, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer tpop.Release()
	major, minor := linux.DecodeDeviceID(dev)
	return t.Kernel().VFS().MknodAt(t, t.Credentials(), &tpop.pop, &vfs.MknodOptions{
		Mode:     linux.FileMode(mode &^ t.FSContext().Umask()),
		DevMajor: uint32(major),
		DevMinor: minor,
	})
}

// Open implements Linux syscall open(2).
func Open(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	addr := args[0].Pointer()
	flags := args[1].Uint()
	mode := args[2].ModeT()
	return openat(t, linux.AT_FDCWD, addr, flags, mode)
}

// Openat implements Linux syscall openat(2).
func Openat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	dirfd := args[0].Int()
	addr := args[1].Pointer()
	flags := args[2].Uint()
	mode := args[3].ModeT()
	return openat(t, dirfd, addr, flags, mode)
}

// Creat implements Linux syscall creat(2).
func Creat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	addr := args[0].Pointer()
	mode := args[1].ModeT()
	return openat(t, linux.AT_FDCWD, addr, linux.O_WRONLY|linux.O_CREAT|linux.O_TRUNC, mode)
}

func openat(t *kernel.Task, dirfd int32, pathAddr usermem.Addr, flags uint32, mode uint) (uintptr, *kernel.SyscallControl, error) {
	path, err := copyInPath(t, pathAddr)
	if err != nil {
		return 0, nil, err
	}
	tpop, err := getTaskPathOperation(t, dirfd, path, disallowEmptyPath, shouldFollowFinalSymlink(flags&linux.O_NOFOLLOW == 0))
	if err != nil {
		return 0, nil, err
	}
	defer tpop.Release()

	file, err := t.Kernel().VFS().OpenAt(t, t.Credentials(), &tpop.pop, &vfs.OpenOptions{
		Flags: flags,
		Mode:  linux.FileMode(mode & (0777 | linux.S_ISUID | linux.S_ISGID | linux.S_ISVTX) &^ t.FSContext().Umask()),
	})
	if err != nil {
		return 0, nil, err
	}
	defer file.DecRef()

	fd, err := t.NewFDFromVFS2(0, file, kernel.FDFlags{
		CloseOnExec: flags&linux.O_CLOEXEC != 0,
	})
	return uintptr(fd), nil, err
}

// Rename implements Linux syscall rename(2).
func Rename(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	oldpathAddr := args[0].Pointer()
	newpathAddr := args[1].Pointer()
	return 0, nil, renameat(t, linux.AT_FDCWD, oldpathAddr, linux.AT_FDCWD, newpathAddr, 0 /* flags */)
}

// Renameat implements Linux syscall renameat(2).
func Renameat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	olddirfd := args[0].Int()
	oldpathAddr := args[1].Pointer()
	newdirfd := args[2].Int()
	newpathAddr := args[3].Pointer()
	return 0, nil, renameat(t, olddirfd, oldpathAddr, newdirfd, newpathAddr, 0 /* flags */)
}

// Renameat2 implements Linux syscall renameat2(2).
func Renameat2(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	olddirfd := args[0].Int()
	oldpathAddr := args[1].Pointer()
	newdirfd := args[2].Int()
	newpathAddr := args[3].Pointer()
	flags := args[4].Uint()
	return 0, nil, renameat(t, olddirfd, oldpathAddr, newdirfd, newpathAddr, flags)
}

func renameat(t *kernel.Task, olddirfd int32, oldpathAddr usermem.Addr, newdirfd int32, newpathAddr usermem.Addr, flags uint32) error {
	oldpath, err := copyInPath(t, oldpathAddr)
	if err != nil {
		return err
	}
	// "If oldpath refers to a symbolic link, the link is renamed" - rename(2)
	oldtpop, err := getTaskPathOperation(t, olddirfd, oldpath, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer oldtpop.Release()

	newpath, err := copyInPath(t, newpathAddr)
	if err != nil {
		return err
	}
	newtpop, err := getTaskPathOperation(t, newdirfd, newpath, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer newtpop.Release()

	return t.Kernel().VFS().RenameAt(t, t.Credentials(), &oldtpop.pop, &newtpop.pop, &vfs.RenameOptions{
		Flags: flags,
	})
}

// Rmdir implements Linux syscall rmdir(2).
func Rmdir(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	pathAddr := args[0].Pointer()
	return 0, nil, rmdirat(t, linux.AT_FDCWD, pathAddr)
}

func rmdirat(t *kernel.Task, dirfd int32, pathAddr usermem.Addr) error {
	path, err := copyInPath(t, pathAddr)
	if err != nil {
		return err
	}
	tpop, err := getTaskPathOperation(t, dirfd, path, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer tpop.Release()
	return t.Kernel().VFS().RmdirAt(t, t.Credentials(), &tpop.pop)
}

// Unlink implements Linux syscall unlink(2).
func Unlink(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	pathAddr := args[0].Pointer()
	return 0, nil, unlinkat(t, linux.AT_FDCWD, pathAddr)
}

func unlinkat(t *kernel.Task, dirfd int32, pathAddr usermem.Addr) error {
	path, err := copyInPath(t, pathAddr)
	if err != nil {
		return err
	}
	tpop, err := getTaskPathOperation(t, dirfd, path, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer tpop.Release()
	return t.Kernel().VFS().UnlinkAt(t, t.Credentials(), &tpop.pop)
}

// Unlinkat implements Linux syscall unlinkat(2).
func Unlinkat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	dirfd := args[0].Int()
	pathAddr := args[1].Pointer()
	flags := args[2].Int()

	if flags&^linux.AT_REMOVEDIR != 0 {
		return 0, nil, syserror.EINVAL
	}

	if flags&linux.AT_REMOVEDIR != 0 {
		return 0, nil, rmdirat(t, dirfd, pathAddr)
	}
	return 0, nil, unlinkat(t, dirfd, pathAddr)
}

// Symlink implements Linux syscall symlink(2).
func Symlink(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	targetAddr := args[0].Pointer()
	linkpathAddr := args[1].Pointer()
	return 0, nil, symlinkat(t, targetAddr, linux.AT_FDCWD, linkpathAddr)
}

// Symlinkat implements Linux syscall symlinkat(2).
func Symlinkat(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
	targetAddr := args[0].Pointer()
	newdirfd := args[1].Int()
	linkpathAddr := args[2].Pointer()
	return 0, nil, symlinkat(t, targetAddr, newdirfd, linkpathAddr)
}

func symlinkat(t *kernel.Task, targetAddr usermem.Addr, newdirfd int32, linkpathAddr usermem.Addr) error {
	target, err := t.CopyInString(targetAddr, linux.PATH_MAX)
	if err != nil {
		return err
	}
	linkpath, err := copyInPath(t, linkpathAddr)
	if err != nil {
		return err
	}
	tpop, err := getTaskPathOperation(t, newdirfd, linkpath, disallowEmptyPath, nofollowFinalSymlink)
	if err != nil {
		return err
	}
	defer tpop.Release()
	return t.Kernel().VFS().SymlinkAt(t, t.Credentials(), &tpop.pop, target)
}