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
|
// SPDX-License-Identifier: GPL-2.0
#define _GNU_SOURCE
#include <errno.h>
#include <fcntl.h>
#include <limits.h>
#include <linux/types.h>
#include <sched.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <syscall.h>
#include <sys/prctl.h>
#include <sys/wait.h>
#include <unistd.h>
#include <sys/socket.h>
#include <sys/stat.h>
#include <linux/kcmp.h>
#include "pidfd.h"
#include "../clone3/clone3_selftests.h"
#include "../kselftest_harness.h"
enum {
PIDFD_NS_USER,
PIDFD_NS_MNT,
PIDFD_NS_PID,
PIDFD_NS_UTS,
PIDFD_NS_IPC,
PIDFD_NS_NET,
PIDFD_NS_CGROUP,
PIDFD_NS_PIDCLD,
PIDFD_NS_MAX
};
const struct ns_info {
const char *name;
int flag;
} ns_info[] = {
[PIDFD_NS_USER] = { "user", CLONE_NEWUSER, },
[PIDFD_NS_MNT] = { "mnt", CLONE_NEWNS, },
[PIDFD_NS_PID] = { "pid", CLONE_NEWPID, },
[PIDFD_NS_UTS] = { "uts", CLONE_NEWUTS, },
[PIDFD_NS_IPC] = { "ipc", CLONE_NEWIPC, },
[PIDFD_NS_NET] = { "net", CLONE_NEWNET, },
[PIDFD_NS_CGROUP] = { "cgroup", CLONE_NEWCGROUP, },
[PIDFD_NS_PIDCLD] = { "pid_for_children", 0, },
};
FIXTURE(current_nsset)
{
pid_t pid;
int pidfd;
int nsfds[PIDFD_NS_MAX];
pid_t child_pid_exited;
int child_pidfd_exited;
pid_t child_pid1;
int child_pidfd1;
int child_nsfds1[PIDFD_NS_MAX];
pid_t child_pid2;
int child_pidfd2;
int child_nsfds2[PIDFD_NS_MAX];
};
static int sys_waitid(int which, pid_t pid, int options)
{
return syscall(__NR_waitid, which, pid, NULL, options, NULL);
}
pid_t create_child(int *pidfd, unsigned flags)
{
struct clone_args args = {
.flags = CLONE_PIDFD | flags,
.exit_signal = SIGCHLD,
.pidfd = ptr_to_u64(pidfd),
};
return sys_clone3(&args, sizeof(struct clone_args));
}
FIXTURE_SETUP(current_nsset)
{
int i, proc_fd, ret;
for (i = 0; i < PIDFD_NS_MAX; i++) {
self->nsfds[i] = -EBADF;
self->child_nsfds1[i] = -EBADF;
self->child_nsfds2[i] = -EBADF;
}
proc_fd = open("/proc/self/ns", O_DIRECTORY | O_CLOEXEC);
ASSERT_GE(proc_fd, 0) {
TH_LOG("%m - Failed to open /proc/self/ns");
}
self->pid = getpid();
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
self->nsfds[i] = openat(proc_fd, info->name, O_RDONLY | O_CLOEXEC);
if (self->nsfds[i] < 0) {
EXPECT_EQ(errno, ENOENT) {
TH_LOG("%m - Failed to open %s namespace for process %d",
info->name, self->pid);
}
}
}
self->pidfd = sys_pidfd_open(self->pid, 0);
EXPECT_GT(self->pidfd, 0) {
TH_LOG("%m - Failed to open pidfd for process %d", self->pid);
}
/* Create task that exits right away. */
self->child_pid_exited = create_child(&self->child_pidfd_exited,
CLONE_NEWUSER | CLONE_NEWNET);
EXPECT_GT(self->child_pid_exited, 0);
if (self->child_pid_exited == 0)
_exit(EXIT_SUCCESS);
ASSERT_EQ(sys_waitid(P_PID, self->child_pid_exited, WEXITED | WNOWAIT), 0);
self->pidfd = sys_pidfd_open(self->pid, 0);
EXPECT_GE(self->pidfd, 0) {
TH_LOG("%m - Failed to open pidfd for process %d", self->pid);
}
/* Create tasks that will be stopped. */
self->child_pid1 = create_child(&self->child_pidfd1,
CLONE_NEWUSER | CLONE_NEWNS |
CLONE_NEWCGROUP | CLONE_NEWIPC |
CLONE_NEWUTS | CLONE_NEWPID |
CLONE_NEWNET);
EXPECT_GE(self->child_pid1, 0);
if (self->child_pid1 == 0) {
pause();
_exit(EXIT_SUCCESS);
}
self->child_pid2 = create_child(&self->child_pidfd2,
CLONE_NEWUSER | CLONE_NEWNS |
CLONE_NEWCGROUP | CLONE_NEWIPC |
CLONE_NEWUTS | CLONE_NEWPID |
CLONE_NEWNET);
EXPECT_GE(self->child_pid2, 0);
if (self->child_pid2 == 0) {
pause();
_exit(EXIT_SUCCESS);
}
for (i = 0; i < PIDFD_NS_MAX; i++) {
char p[100];
const struct ns_info *info = &ns_info[i];
self->nsfds[i] = openat(proc_fd, info->name, O_RDONLY | O_CLOEXEC);
if (self->nsfds[i] < 0) {
EXPECT_EQ(errno, ENOENT) {
TH_LOG("%m - Failed to open %s namespace for process %d",
info->name, self->pid);
}
}
ret = snprintf(p, sizeof(p), "/proc/%d/ns/%s",
self->child_pid1, info->name);
EXPECT_GT(ret, 0);
EXPECT_LT(ret, sizeof(p));
self->child_nsfds1[i] = open(p, O_RDONLY | O_CLOEXEC);
if (self->child_nsfds1[i] < 0) {
EXPECT_EQ(errno, ENOENT) {
TH_LOG("%m - Failed to open %s namespace for process %d",
info->name, self->child_pid1);
}
}
ret = snprintf(p, sizeof(p), "/proc/%d/ns/%s",
self->child_pid2, info->name);
EXPECT_GT(ret, 0);
EXPECT_LT(ret, sizeof(p));
self->child_nsfds2[i] = open(p, O_RDONLY | O_CLOEXEC);
if (self->child_nsfds2[i] < 0) {
EXPECT_EQ(errno, ENOENT) {
TH_LOG("%m - Failed to open %s namespace for process %d",
info->name, self->child_pid1);
}
}
}
close(proc_fd);
}
FIXTURE_TEARDOWN(current_nsset)
{
int i;
ASSERT_EQ(sys_pidfd_send_signal(self->child_pidfd1,
SIGKILL, NULL, 0), 0);
ASSERT_EQ(sys_pidfd_send_signal(self->child_pidfd2,
SIGKILL, NULL, 0), 0);
for (i = 0; i < PIDFD_NS_MAX; i++) {
if (self->nsfds[i] >= 0)
close(self->nsfds[i]);
if (self->child_nsfds1[i] >= 0)
close(self->child_nsfds1[i]);
if (self->child_nsfds2[i] >= 0)
close(self->child_nsfds2[i]);
}
if (self->child_pidfd1 >= 0)
EXPECT_EQ(0, close(self->child_pidfd1));
if (self->child_pidfd2 >= 0)
EXPECT_EQ(0, close(self->child_pidfd2));
ASSERT_EQ(sys_waitid(P_PID, self->child_pid_exited, WEXITED), 0);
ASSERT_EQ(sys_waitid(P_PID, self->child_pid1, WEXITED), 0);
ASSERT_EQ(sys_waitid(P_PID, self->child_pid2, WEXITED), 0);
}
static int preserve_ns(const int pid, const char *ns)
{
int ret;
char path[50];
ret = snprintf(path, sizeof(path), "/proc/%d/ns/%s", pid, ns);
if (ret < 0 || (size_t)ret >= sizeof(path))
return -EIO;
return open(path, O_RDONLY | O_CLOEXEC);
}
static int in_same_namespace(int ns_fd1, pid_t pid2, const char *ns)
{
int ns_fd2 = -EBADF;
int ret = -1;
struct stat ns_st1, ns_st2;
ret = fstat(ns_fd1, &ns_st1);
if (ret < 0)
return -1;
ns_fd2 = preserve_ns(pid2, ns);
if (ns_fd2 < 0)
return -1;
ret = fstat(ns_fd2, &ns_st2);
close(ns_fd2);
if (ret < 0)
return -1;
/* processes are in the same namespace */
if ((ns_st1.st_dev == ns_st2.st_dev) &&
(ns_st1.st_ino == ns_st2.st_ino))
return 1;
/* processes are in different namespaces */
return 0;
}
/* Test that we can't pass garbage to the kernel. */
TEST_F(current_nsset, invalid_flags)
{
ASSERT_NE(setns(self->pidfd, 0), 0);
EXPECT_EQ(errno, EINVAL);
ASSERT_NE(setns(self->pidfd, -1), 0);
EXPECT_EQ(errno, EINVAL);
ASSERT_NE(setns(self->pidfd, CLONE_VM), 0);
EXPECT_EQ(errno, EINVAL);
ASSERT_NE(setns(self->pidfd, CLONE_NEWUSER | CLONE_VM), 0);
EXPECT_EQ(errno, EINVAL);
}
/* Test that we can't attach to a task that has already exited. */
TEST_F(current_nsset, pidfd_exited_child)
{
int i;
pid_t pid;
ASSERT_NE(setns(self->child_pidfd_exited, CLONE_NEWUSER | CLONE_NEWNET),
0);
EXPECT_EQ(errno, ESRCH);
pid = getpid();
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
/* Verify that we haven't changed any namespaces. */
if (self->nsfds[i] >= 0)
ASSERT_EQ(in_same_namespace(self->nsfds[i], pid, info->name), 1);
}
}
TEST_F(current_nsset, pidfd_incremental_setns)
{
int i;
pid_t pid;
pid = getpid();
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
int nsfd;
if (self->child_nsfds1[i] < 0)
continue;
if (info->flag) {
ASSERT_EQ(setns(self->child_pidfd1, info->flag), 0) {
TH_LOG("%m - Failed to setns to %s namespace of %d via pidfd %d",
info->name, self->child_pid1,
self->child_pidfd1);
}
}
/* Verify that we have changed to the correct namespaces. */
if (info->flag == CLONE_NEWPID)
nsfd = self->nsfds[i];
else
nsfd = self->child_nsfds1[i];
ASSERT_EQ(in_same_namespace(nsfd, pid, info->name), 1) {
TH_LOG("setns failed to place us correctly into %s namespace of %d via pidfd %d",
info->name, self->child_pid1,
self->child_pidfd1);
}
TH_LOG("Managed to correctly setns to %s namespace of %d via pidfd %d",
info->name, self->child_pid1, self->child_pidfd1);
}
}
TEST_F(current_nsset, nsfd_incremental_setns)
{
int i;
pid_t pid;
pid = getpid();
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
int nsfd;
if (self->child_nsfds1[i] < 0)
continue;
if (info->flag) {
ASSERT_EQ(setns(self->child_nsfds1[i], info->flag), 0) {
TH_LOG("%m - Failed to setns to %s namespace of %d via nsfd %d",
info->name, self->child_pid1,
self->child_nsfds1[i]);
}
}
/* Verify that we have changed to the correct namespaces. */
if (info->flag == CLONE_NEWPID)
nsfd = self->nsfds[i];
else
nsfd = self->child_nsfds1[i];
ASSERT_EQ(in_same_namespace(nsfd, pid, info->name), 1) {
TH_LOG("setns failed to place us correctly into %s namespace of %d via nsfd %d",
info->name, self->child_pid1,
self->child_nsfds1[i]);
}
TH_LOG("Managed to correctly setns to %s namespace of %d via nsfd %d",
info->name, self->child_pid1, self->child_nsfds1[i]);
}
}
TEST_F(current_nsset, pidfd_one_shot_setns)
{
unsigned flags = 0;
int i;
pid_t pid;
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
if (self->child_nsfds1[i] < 0)
continue;
flags |= info->flag;
TH_LOG("Adding %s namespace of %d to list of namespaces to attach to",
info->name, self->child_pid1);
}
ASSERT_EQ(setns(self->child_pidfd1, flags), 0) {
TH_LOG("%m - Failed to setns to namespaces of %d",
self->child_pid1);
}
pid = getpid();
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
int nsfd;
if (self->child_nsfds1[i] < 0)
continue;
/* Verify that we have changed to the correct namespaces. */
if (info->flag == CLONE_NEWPID)
nsfd = self->nsfds[i];
else
nsfd = self->child_nsfds1[i];
ASSERT_EQ(in_same_namespace(nsfd, pid, info->name), 1) {
TH_LOG("setns failed to place us correctly into %s namespace of %d",
info->name, self->child_pid1);
}
TH_LOG("Managed to correctly setns to %s namespace of %d",
info->name, self->child_pid1);
}
}
TEST_F(current_nsset, no_foul_play)
{
unsigned flags = 0;
int i;
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
if (self->child_nsfds1[i] < 0)
continue;
flags |= info->flag;
if (info->flag) /* No use logging pid_for_children. */
TH_LOG("Adding %s namespace of %d to list of namespaces to attach to",
info->name, self->child_pid1);
}
ASSERT_EQ(setns(self->child_pidfd1, flags), 0) {
TH_LOG("%m - Failed to setns to namespaces of %d vid pidfd %d",
self->child_pid1, self->child_pidfd1);
}
/*
* Can't setns to a user namespace outside of our hierarchy since we
* don't have caps in there and didn't create it. That means that under
* no circumstances should we be able to setns to any of the other
* ones since they aren't owned by our user namespace.
*/
for (i = 0; i < PIDFD_NS_MAX; i++) {
const struct ns_info *info = &ns_info[i];
if (self->child_nsfds2[i] < 0 || !info->flag)
continue;
ASSERT_NE(setns(self->child_pidfd2, info->flag), 0) {
TH_LOG("Managed to setns to %s namespace of %d via pidfd %d",
info->name, self->child_pid2,
self->child_pidfd2);
}
TH_LOG("%m - Correctly failed to setns to %s namespace of %d via pidfd %d",
info->name, self->child_pid2,
self->child_pidfd2);
ASSERT_NE(setns(self->child_nsfds2[i], info->flag), 0) {
TH_LOG("Managed to setns to %s namespace of %d via nsfd %d",
info->name, self->child_pid2,
self->child_nsfds2[i]);
}
TH_LOG("%m - Correctly failed to setns to %s namespace of %d via nsfd %d",
info->name, self->child_pid2,
self->child_nsfds2[i]);
}
}
TEST(setns_einval)
{
int fd;
fd = sys_memfd_create("rostock", 0);
EXPECT_GT(fd, 0);
ASSERT_NE(setns(fd, 0), 0);
EXPECT_EQ(errno, EINVAL);
close(fd);
}
TEST_HARNESS_MAIN
|