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
|
#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include <fcntl.h>
#include <limits.h>
#include <sched.h>
#include <unistd.h>
#include <sys/ioctl.h>
#include <sys/wait.h>
#include <linux/seccomp.h>
#include <linux/filter.h>
#include <linux/audit.h>
#include <errno.h>
#include "gluten.h"
#include "operations.h"
static bool is_cookie_valid(int notifyFd, uint64_t id)
{
return ioctl(notifyFd, SECCOMP_IOCTL_NOTIF_ID_VALID, &id) == 0;
}
static int send_target(const struct seccomp_notif_resp *resp, int notifyfd)
{
if (!is_cookie_valid(notifyfd, resp->id)) {
fprintf(stderr,
"the response id isn't valid\ncheck if the targets has already terminated\n");
return -1;
}
if (ioctl(notifyfd, SECCOMP_IOCTL_NOTIF_SEND, resp) < 0) {
if (errno != EINPROGRESS) {
perror("sending the response");
return -1;
}
}
return 0;
}
static int send_inject_target(const struct seccomp_notif_addfd *resp,
int notifyfd)
{
if (!is_cookie_valid(notifyfd, resp->id)) {
fprintf(stderr,
"the response id isn't valid\ncheck if the targets has already terminated\n");
return -1;
}
if (ioctl(notifyfd, SECCOMP_IOCTL_NOTIF_ADDFD, resp) < 0) {
if (errno != EINPROGRESS) {
perror("sending the response");
return -1;
}
}
return 0;
}
static void proc_ns_name(unsigned i, char *ns)
{
switch (i) {
case NS_CGROUP:
snprintf(ns, PATH_MAX + 1, "cgroup");
break;
case NS_IPC:
snprintf(ns, PATH_MAX + 1, "ipc");
break;
case NS_NET:
snprintf(ns, PATH_MAX + 1, "net");
break;
case NS_MOUNT:
snprintf(ns, PATH_MAX + 1, "mnt");
break;
case NS_PID:
snprintf(ns, PATH_MAX + 1, "pid");
break;
case NS_USER:
snprintf(ns, PATH_MAX + 1, "user");
break;
case NS_UTS:
snprintf(ns, PATH_MAX + 1, "uts");
break;
case NS_TIME:
snprintf(ns, PATH_MAX + 1, "time");
break;
default:
fprintf(stderr, "unrecognized namespace index %d\n", i);
}
}
static int set_namespaces(const struct op_call *a, int tpid)
{
char path[PATH_MAX + 1];
char ns_name[PATH_MAX / 2];
struct ns_spec ns;
int fd;
unsigned int i;
for (i = 0, ns = (a->context).ns[i]; i < sizeof(enum ns_type);
i++, ns = (a->context).ns[i]) {
proc_ns_name(i, ns_name);
switch (ns.type) {
case NS_NONE:
continue;
case NS_SPEC_TARGET:
snprintf(path, sizeof(path), "/proc/%d/ns/%s", tpid,
ns_name);
break;
case NS_SPEC_PID:
snprintf(path, sizeof(path), "/proc/%d/ns/%s", ns.pid,
ns_name);
break;
case NS_SPEC_PATH:
snprintf(path, sizeof(path), "%s", ns.path);
break;
}
if ((fd = open(path, O_CLOEXEC)) < 0) {
fprintf(stderr, "open for file %s: %s", path,
strerror(errno));
return -1;
}
if (setns(fd, 0) != 0) {
perror("setns");
return -1;
}
}
return 0;
}
static int execute_syscall(void *args)
{
struct arg_clone *a = (struct arg_clone *)args;
const struct op_call *c = a->args;
if (set_namespaces(a->args, a->pid) < 0) {
exit(EXIT_FAILURE);
}
/* execute syscall */
a->ret = syscall(c->nr, c->args[0], c->args[1], c->args[2], c->args[3],
c->args[4], c->args[5]);
a->err = errno;
if (a->ret < 0) {
perror("syscall");
exit(EXIT_FAILURE);
}
exit(0);
}
int copy_args(struct seccomp_notif *req, struct op_copy_args *copy, void *data,
int notifier)
{
char path[PATH_MAX];
unsigned int i;
ssize_t nread;
void *dest;
int fd;
snprintf(path, sizeof(path), "/proc/%d/mem", req->pid);
if ((fd = open(path, O_RDONLY | O_CLOEXEC)) < 0) {
perror("open mem");
return -1;
}
/*
* Avoid the TOCTOU and check if the read mappings are still valid
*/
if (!is_cookie_valid(notifier, req->id)) {
fprintf(stderr, "the seccomp request isn't valid anymore\n");
return -1;
}
for (i = 0; i < 6; i++) {
if (copy->args[i].type == REFERENCE) {
dest = (uint16_t *)data + copy->args[i].args_off;
nread = pread(fd, dest, copy->args[i].size,
req->data.args[i]);
if (nread < 0) {
perror("pread");
return -1;
}
} else {
memcpy((uint16_t *)data + copy->args[i].args_off,
&req->data.args[i], copy->args[i].size);
}
}
close(fd);
return 0;
}
int do_call(struct arg_clone *c)
{
char stack[STACK_SIZE];
pid_t child;
c->ret = -1;
c->err = 0;
/* Create a process that will be moved to the namespace */
child = clone(execute_syscall, stack + sizeof(stack),
CLONE_FILES | CLONE_VM | SIGCHLD, (void *)c);
if (child == -1) {
perror("clone");
return -1;
}
wait(NULL);
return 0;
}
static void set_inject_fields(uint64_t id, void *data, const struct op *a,
struct seccomp_notif_addfd *resp)
{
const struct fd_type *new = &(a->inj).newfd;
const struct fd_type *old = &(a->inj).oldfd;
resp->flags = SECCOMP_ADDFD_FLAG_SETFD;
resp->id = id;
if (new->type == IMMEDIATE)
resp->newfd = new->fd;
else
memcpy(&resp->srcfd, (uint16_t *)data + old->fd_off,
sizeof(resp->srcfd));
if (old->type == IMMEDIATE)
resp->srcfd = old->fd;
else
memcpy(&resp->srcfd, (uint16_t *)data + old->fd_off,
sizeof(resp->srcfd));
resp->newfd_flags = 0;
}
int do_operations(void *data, struct op operations[], struct seccomp_notif *req,
unsigned int n_operations, int pid, int notifyfd, uint64_t id)
{
struct seccomp_notif_addfd resp_fd;
struct seccomp_notif_resp resp;
struct arg_clone c;
unsigned int i;
for (i = 0; i < n_operations; i++) {
switch (operations[i].type) {
case OP_CALL:
resp.id = id;
resp.val = 0;
resp.flags = 0;
c.args = &operations[i].call;
c.pid = pid;
if (do_call(&c) == -1) {
resp.error = -1;
if (send_target(&resp, notifyfd) == -1)
return -1;
}
if (c.err != 0) {
resp.error = -1;
if (send_target(&resp, notifyfd) == -1)
return c.err;
}
/*
* The result of the call needs to be save as
* reference
*/
if (operations[i].call.has_ret) {
memcpy((uint16_t *)data +
operations[i].call.ret_off,
&c.ret, sizeof(c.ret));
}
break;
case OP_BLOCK:
resp.id = id;
resp.val = 0;
resp.flags = 0;
resp.error = operations[i].block.error;
if (send_target(&resp, notifyfd) == -1)
return -1;
break;
case OP_RETURN:
resp.id = id;
resp.flags = 0;
resp.error = 0;
if (operations[i].ret.type == IMMEDIATE)
resp.val = operations[i].ret.value;
else
memcpy(&resp.val,
(uint16_t *)data +
operations[i].ret.value_off,
sizeof(resp.val));
if (send_target(&resp, notifyfd) == -1)
return -1;
break;
case OP_CONT:
resp.id = id;
resp.flags = SECCOMP_USER_NOTIF_FLAG_CONTINUE;
resp.error = 0;
resp.val = 0;
if (send_target(&resp, notifyfd) == -1)
return -1;
break;
case OP_INJECT_A:
set_inject_fields(id, data, &operations[i], &resp_fd);
resp_fd.flags |= SECCOMP_ADDFD_FLAG_SEND;
if (send_inject_target(&resp_fd, notifyfd) == -1)
return -1;
break;
case OP_INJECT:
set_inject_fields(id, data, &operations[i], &resp_fd);
if (send_inject_target(&resp_fd, notifyfd) == -1)
return -1;
break;
case OP_COPY_ARGS:
if (copy_args(req, &operations[i].copy, data,
notifyfd) < 0)
return -1;
break;
default:
fprintf(stderr, "unknow operation %d \n",
operations[i].type);
}
}
return 0;
}
|