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
|
// 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/fs/lock"
"gvisor.dev/gvisor/pkg/sentry/fsimpl/tmpfs"
"gvisor.dev/gvisor/pkg/sentry/kernel"
"gvisor.dev/gvisor/pkg/sentry/kernel/pipe"
slinux "gvisor.dev/gvisor/pkg/sentry/syscalls/linux"
"gvisor.dev/gvisor/pkg/sentry/vfs"
"gvisor.dev/gvisor/pkg/syserror"
)
// Close implements Linux syscall close(2).
func Close(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
fd := args[0].Int()
// Note that Remove provides a reference on the file that we may use to
// flush. It is still active until we drop the final reference below
// (and other reference-holding operations complete).
_, file := t.FDTable().Remove(fd)
if file == nil {
return 0, nil, syserror.EBADF
}
defer file.DecRef()
err := file.OnClose(t)
return 0, nil, slinux.HandleIOErrorVFS2(t, false /* partial */, err, syserror.EINTR, "close", file)
}
// Dup implements Linux syscall dup(2).
func Dup(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
fd := args[0].Int()
file := t.GetFileVFS2(fd)
if file == nil {
return 0, nil, syserror.EBADF
}
defer file.DecRef()
newFD, err := t.NewFDFromVFS2(0, file, kernel.FDFlags{})
if err != nil {
return 0, nil, syserror.EMFILE
}
return uintptr(newFD), nil, nil
}
// Dup2 implements Linux syscall dup2(2).
func Dup2(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
oldfd := args[0].Int()
newfd := args[1].Int()
if oldfd == newfd {
// As long as oldfd is valid, dup2() does nothing and returns newfd.
file := t.GetFileVFS2(oldfd)
if file == nil {
return 0, nil, syserror.EBADF
}
file.DecRef()
return uintptr(newfd), nil, nil
}
return dup3(t, oldfd, newfd, 0)
}
// Dup3 implements Linux syscall dup3(2).
func Dup3(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
oldfd := args[0].Int()
newfd := args[1].Int()
flags := args[2].Uint()
if oldfd == newfd {
return 0, nil, syserror.EINVAL
}
return dup3(t, oldfd, newfd, flags)
}
func dup3(t *kernel.Task, oldfd, newfd int32, flags uint32) (uintptr, *kernel.SyscallControl, error) {
if flags&^linux.O_CLOEXEC != 0 {
return 0, nil, syserror.EINVAL
}
file := t.GetFileVFS2(oldfd)
if file == nil {
return 0, nil, syserror.EBADF
}
defer file.DecRef()
err := t.NewFDAtVFS2(newfd, file, kernel.FDFlags{
CloseOnExec: flags&linux.O_CLOEXEC != 0,
})
if err != nil {
return 0, nil, err
}
return uintptr(newfd), nil, nil
}
// Fcntl implements linux syscall fcntl(2).
func Fcntl(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
fd := args[0].Int()
cmd := args[1].Int()
file, flags := t.FDTable().GetVFS2(fd)
if file == nil {
return 0, nil, syserror.EBADF
}
defer file.DecRef()
switch cmd {
case linux.F_DUPFD, linux.F_DUPFD_CLOEXEC:
minfd := args[2].Int()
fd, err := t.NewFDFromVFS2(minfd, file, kernel.FDFlags{
CloseOnExec: cmd == linux.F_DUPFD_CLOEXEC,
})
if err != nil {
return 0, nil, err
}
return uintptr(fd), nil, nil
case linux.F_GETFD:
return uintptr(flags.ToLinuxFDFlags()), nil, nil
case linux.F_SETFD:
flags := args[2].Uint()
err := t.FDTable().SetFlagsVFS2(fd, kernel.FDFlags{
CloseOnExec: flags&linux.FD_CLOEXEC != 0,
})
return 0, nil, err
case linux.F_GETFL:
return uintptr(file.StatusFlags()), nil, nil
case linux.F_SETFL:
return 0, nil, file.SetStatusFlags(t, t.Credentials(), args[2].Uint())
case linux.F_SETPIPE_SZ:
pipefile, ok := file.Impl().(*pipe.VFSPipeFD)
if !ok {
return 0, nil, syserror.EBADF
}
n, err := pipefile.SetPipeSize(int64(args[2].Int()))
if err != nil {
return 0, nil, err
}
return uintptr(n), nil, nil
case linux.F_GETPIPE_SZ:
pipefile, ok := file.Impl().(*pipe.VFSPipeFD)
if !ok {
return 0, nil, syserror.EBADF
}
return uintptr(pipefile.PipeSize()), nil, nil
case linux.F_GET_SEALS:
val, err := tmpfs.GetSeals(file)
return uintptr(val), nil, err
case linux.F_ADD_SEALS:
if !file.IsWritable() {
return 0, nil, syserror.EPERM
}
err := tmpfs.AddSeals(file, args[2].Uint())
return 0, nil, err
case linux.F_SETLK, linux.F_SETLKW:
return 0, nil, posixLock(t, args, file, cmd)
default:
// TODO(gvisor.dev/issue/2920): Everything else is not yet supported.
return 0, nil, syserror.EINVAL
}
}
func posixLock(t *kernel.Task, args arch.SyscallArguments, file *vfs.FileDescription, cmd int32) error {
// Copy in the lock request.
flockAddr := args[2].Pointer()
var flock linux.Flock
if _, err := t.CopyIn(flockAddr, &flock); err != nil {
return err
}
var blocker lock.Blocker
if cmd == linux.F_SETLKW {
blocker = t
}
switch flock.Type {
case linux.F_RDLCK:
if !file.IsReadable() {
return syserror.EBADF
}
return file.LockPOSIX(t, t.FDTable(), lock.ReadLock, uint64(flock.Start), uint64(flock.Len), flock.Whence, blocker)
case linux.F_WRLCK:
if !file.IsWritable() {
return syserror.EBADF
}
return file.LockPOSIX(t, t.FDTable(), lock.WriteLock, uint64(flock.Start), uint64(flock.Len), flock.Whence, blocker)
case linux.F_UNLCK:
return file.UnlockPOSIX(t, t.FDTable(), uint64(flock.Start), uint64(flock.Len), flock.Whence)
default:
return syserror.EINVAL
}
}
// Fadvise64 implements fadvise64(2).
// This implementation currently ignores the provided advice.
func Fadvise64(t *kernel.Task, args arch.SyscallArguments) (uintptr, *kernel.SyscallControl, error) {
fd := args[0].Int()
length := args[2].Int64()
advice := args[3].Int()
// Note: offset is allowed to be negative.
if length < 0 {
return 0, nil, syserror.EINVAL
}
file := t.GetFileVFS2(fd)
if file == nil {
return 0, nil, syserror.EBADF
}
defer file.DecRef()
// If the FD refers to a pipe or FIFO, return error.
if _, isPipe := file.Impl().(*pipe.VFSPipeFD); isPipe {
return 0, nil, syserror.ESPIPE
}
switch advice {
case linux.POSIX_FADV_NORMAL:
case linux.POSIX_FADV_RANDOM:
case linux.POSIX_FADV_SEQUENTIAL:
case linux.POSIX_FADV_WILLNEED:
case linux.POSIX_FADV_DONTNEED:
case linux.POSIX_FADV_NOREUSE:
default:
return 0, nil, syserror.EINVAL
}
// Sure, whatever.
return 0, nil, nil
}
|
