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
|
/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4; fill-column: 100 -*- */
/*
* This file is part of the LibreOffice project.
*
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/.
*/
/*
* Code to lock-down the environment of the processes we run, to avoid
* exotic or un-necessary system calls to be used to break containment.
*/
#include <config.h>
#include "Seccomp.hpp"
#include <dlfcn.h>
#include <ftw.h>
#include <linux/audit.h>
#include <linux/filter.h>
#if DISABLE_SECCOMP == 0
#include <linux/seccomp.h>
#endif
#include <malloc.h>
#include <signal.h>
#include <sys/capability.h>
#include <sys/prctl.h>
#include <sys/resource.h>
#include <sys/syscall.h>
#include <sys/time.h>
#include <unistd.h>
#include <utime.h>
#include <common/Log.hpp>
#include <common/SigUtil.hpp>
#if DISABLE_SECCOMP == 0
#ifndef SYS_SECCOMP
# define SYS_SECCOMP 1
#endif
#if defined(__x86_64__)
# define AUDIT_ARCH_NR AUDIT_ARCH_X86_64
# define SECCOMP_REG(_ctx, _reg) ((_ctx)->uc_mcontext.gregs[(_reg)])
# define SECCOMP_SYSCALL(_ctx) SECCOMP_REG(_ctx, REG_RAX)
#elif defined(__arm__)
# define AUDIT_ARCH_NR AUDIT_ARCH_ARM
# define SECCOMP_REG(_ctx, _reg) ((_ctx)->uc_mcontext.arm_##_reg)
# define SECCOMP_SYSCALL(_ctx) SECCOMP_REG(_ctx, r7)
#else
# error "Platform does not support seccomp filtering yet - unsafe."
#endif
extern "C" {
static void handleSysSignal(int /* signal */,
siginfo_t *info,
void *context)
{
ucontext_t *uctx = static_cast<ucontext_t *>(context);
Log::signalLogPrefix();
Log::signalLog("SIGSYS trapped with code: ");
Log::signalLogNumber(info->si_code);
Log::signalLog(" and context ");
Log::signalLogNumber(reinterpret_cast<size_t>(context));
Log::signalLog("\n");
if (info->si_code != SYS_SECCOMP || !uctx)
return;
unsigned int syscall = SECCOMP_SYSCALL (uctx);
Log::signalLogPrefix();
Log::signalLog(" seccomp trapped signal, un-authorized sys-call: ");
Log::signalLogNumber(syscall);
Log::signalLog("\n");
SigUtil::dumpBacktrace();
Log::shutdown();
_exit(1);
}
} // extern "C"
#endif
namespace Seccomp {
bool lockdown(Type type)
{
(void)type; // so far just the kit.
#if DISABLE_SECCOMP == 0
#define ACCEPT_SYSCALL(name) \
BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, __NR_##name, 0, 1), \
BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_ALLOW)
#define KILL_SYSCALL_FULL(fullname) \
BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, fullname, 0, 1), \
BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_TRAP)
#define KILL_SYSCALL(name) \
KILL_SYSCALL_FULL(__NR_##name)
struct sock_filter filterCode[] = {
// Check our architecture is correct.
BPF_STMT(BPF_LD+BPF_W+BPF_ABS, offsetof(struct seccomp_data, arch)),
BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, AUDIT_ARCH_NR, 1, 0),
BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_KILL),
// Load syscall number
BPF_STMT(BPF_LD+BPF_W+BPF_ABS, offsetof(struct seccomp_data, nr)),
// First white-list the syscalls we frequently use.
ACCEPT_SYSCALL(recvfrom),
ACCEPT_SYSCALL(write),
ACCEPT_SYSCALL(futex),
// glibc's 'poll' has to answer for this lot:
ACCEPT_SYSCALL(epoll_wait),
ACCEPT_SYSCALL(epoll_ctl),
ACCEPT_SYSCALL(epoll_create),
ACCEPT_SYSCALL(close),
ACCEPT_SYSCALL(nanosleep),
// Now block everything that we don't like the look of.
// FIXME: should we bother blocking calls that have early
// permission checks we don't meet ?
#if 0
// cf. eg. /usr/include/asm/unistd_64.h ...
KILL_SYSCALL(ioctl),
KILL_SYSCALL(mincore),
KILL_SYSCALL(shmget),
KILL_SYSCALL(shmat),
KILL_SYSCALL(shmctl),
#endif
KILL_SYSCALL(getitimer),
KILL_SYSCALL(setitimer),
KILL_SYSCALL(sendfile),
KILL_SYSCALL(listen), // server sockets
KILL_SYSCALL(accept), // server sockets
#if 0
KILL_SYSCALL(wait4),
#endif
KILL_SYSCALL(kill), // !
KILL_SYSCALL(shmctl),
KILL_SYSCALL(ptrace), // tracing
KILL_SYSCALL(capset),
KILL_SYSCALL(uselib),
KILL_SYSCALL(personality), // !
KILL_SYSCALL(vhangup),
#ifdef __NR_modify_ldt
KILL_SYSCALL(modify_ldt), // !
#endif
#ifdef __PNR_modify_ldt
KILL_SYSCALL_FULL(__PNR_modify_ldt), // !
#endif
KILL_SYSCALL(pivot_root), // !
KILL_SYSCALL(chroot),
KILL_SYSCALL(acct), // !
KILL_SYSCALL(sync), // I/O perf.
KILL_SYSCALL(mount),
KILL_SYSCALL(umount2),
KILL_SYSCALL(swapon),
KILL_SYSCALL(swapoff),
KILL_SYSCALL(reboot), // !
KILL_SYSCALL(sethostname),
KILL_SYSCALL(setdomainname),
KILL_SYSCALL(tkill),
KILL_SYSCALL(mbind), // vm bits
KILL_SYSCALL(set_mempolicy), // vm bits
KILL_SYSCALL(get_mempolicy), // vm bits
KILL_SYSCALL(kexec_load),
KILL_SYSCALL(add_key), // kernel keyring
KILL_SYSCALL(request_key), // kernel keyring
KILL_SYSCALL(keyctl), // kernel keyring
KILL_SYSCALL(inotify_init),
KILL_SYSCALL(inotify_add_watch),
KILL_SYSCALL(inotify_rm_watch),
KILL_SYSCALL(unshare),
KILL_SYSCALL(splice),
KILL_SYSCALL(tee),
KILL_SYSCALL(vmsplice), // vm bits
KILL_SYSCALL(move_pages), // vm bits
KILL_SYSCALL(accept4), // server sockets
KILL_SYSCALL(inotify_init1),
KILL_SYSCALL(perf_event_open), // profiling
KILL_SYSCALL(fanotify_init),
KILL_SYSCALL(fanotify_mark),
#ifdef __NR_seccomp
KILL_SYSCALL(seccomp), // no further fiddling
#endif
#ifdef __NR_bpf
KILL_SYSCALL(bpf), // no further fiddling
#endif
// allow the rest.
BPF_STMT(BPF_RET+BPF_K, SECCOMP_RET_ALLOW)
};
struct sock_fprog filter = {
sizeof(filterCode)/sizeof(filterCode[0]), // length
filterCode
};
if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0))
{
LOG_ERR("Cannot turn off acquisition of new privileges for us & children");
return false;
}
if (prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, &filter))
{
LOG_ERR("Failed to install seccomp syscall filter");
return false;
}
// Trap, log, and exit on failure
struct sigaction action;
sigemptyset(&action.sa_mask);
action.sa_flags = SA_SIGINFO;
action.sa_sigaction = handleSysSignal;
sigaction(SIGSYS, &action, nullptr);
LOG_TRC("Install seccomp filter successfully.");
return true;
#else // DISABLE_SECCOMP == 0
LOG_WRN("Warning: this binary was compiled with disabled seccomp-bpf.");
return true;
#endif // DISABLE_SECCOMP == 0
}
} // namespace Seccomp
namespace Rlimit {
void setRLimit(rlim_t confLim, int resource, const std::string &resourceText, const std::string &unitText)
{
rlim_t lim = confLim;
if (lim <= 0)
lim = RLIM_INFINITY;
const std::string limTextWithUnit((lim == RLIM_INFINITY) ? "unlimited" : std::to_string(lim) + " " + unitText);
if (resource != RLIMIT_FSIZE && resource != RLIMIT_NOFILE)
{
/* FIXME Currently the RLIMIT_FSIZE handling is non-ideal, and can
* lead to crashes of the kit processes due to not handling signal
* 25 gracefully. Let's disable for now before there's a more
* concrete plan.
* Similar issues with RLIMIT_NOFILE
*/
rlimit rlim = { lim, lim };
if (setrlimit(resource, &rlim) != 0)
LOG_SYS("Failed to set " << resourceText << " to " << limTextWithUnit << ".");
if (getrlimit(resource, &rlim) == 0)
{
const std::string setLimTextWithUnit((rlim.rlim_max == RLIM_INFINITY) ? "unlimited" : std::to_string(rlim.rlim_max) + " " + unitText);
LOG_INF(resourceText << " is " << setLimTextWithUnit << " after setting it to " << limTextWithUnit << ".");
}
else
LOG_SYS("Failed to get " << resourceText << ".");
}
else
LOG_INF("Ignored setting " << resourceText << " to " << limTextWithUnit << ".");
}
bool handleSetrlimitCommand(const std::vector<std::string>& tokens)
{
if (tokens.size() == 3 && tokens[0] == "setconfig")
{
if (tokens[1] == "limit_virt_mem_mb")
{
setRLimit(std::stoi(tokens[2]) * 1024 * 1024, RLIMIT_AS, "RLIMIT_AS", "bytes");
}
else if (tokens[1] == "limit_stack_mem_kb")
{
setRLimit(std::stoi(tokens[2]) * 1024, RLIMIT_STACK, "RLIMIT_STACK", "bytes");
}
else if (tokens[1] == "limit_file_size_mb")
{
setRLimit(std::stoi(tokens[2]) * 1024 * 1024, RLIMIT_FSIZE, "RLIMIT_FSIZE", "bytes");
}
else if (tokens[1] == "limit_num_open_files")
{
setRLimit(std::stoi(tokens[2]), RLIMIT_NOFILE, "RLIMIT_NOFILE", "files");
}
else
return false;
return true;
}
return false;
}
} // namespace Rlimit
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */
|