1SECCOMP(2) Linux Programmer's Manual SECCOMP(2)
2
6 seccomp - operate on Secure Computing state of the process
7
9 #include <linux/seccomp.h>
10 #include <linux/filter.h>
11 #include <linux/audit.h>
12 #include <linux/signal.h>
13 #include <sys/ptrace.h>
14 int seccomp(unsigned int operation, unsigned int flags, void *args);
16
18 The seccomp() system call operates on the Secure Computing (seccomp)
19 state of the calling process.
20 Currently, Linux supports the following operation values:
22 SECCOMP_SET_MODE_STRICT
24 The only system calls that the calling thread is permitted to
25 make are read(2), write(2), _exit(2) (but not exit_group(2)),
26 and sigreturn(2). Other system calls result in the delivery of
27 a SIGKILL signal. Strict secure computing mode is useful for
28 number-crunching applications that may need to execute untrusted
29 byte code, perhaps obtained by reading from a pipe or socket.
30 Note that although the calling thread can no longer call sig‐
32 procmask(2), it can use sigreturn(2) to block all signals apart
33 from SIGKILL and SIGSTOP. This means that alarm(2) (for exam‐
34 ple) is not sufficient for restricting the process's execution
35 time. Instead, to reliably terminate the process, SIGKILL must
36 be used. This can be done by using timer_create(2) with
37 SIGEV_SIGNAL and sigev_signo set to SIGKILL, or by using setr‐
38 limit(2) to set the hard limit for RLIMIT_CPU.
39 This operation is available only if the kernel is configured
41 with CONFIG_SECCOMP enabled.
42 The value of flags must be 0, and args must be NULL.
44 This operation is functionally identical to the call:
46 prctl(PR_SET_SECCOMP, SECCOMP_MODE_STRICT);
48 SECCOMP_SET_MODE_FILTER
50 The system calls allowed are defined by a pointer to a Berkeley
51 Packet Filter (BPF) passed via args. This argument is a pointer
52 to a struct sock_fprog; it can be designed to filter arbitrary
53 system calls and system call arguments. If the filter is
54 invalid, seccomp() fails, returning EINVAL in errno.
55 If fork(2) or clone(2) is allowed by the filter, any child pro‐
57 cesses will be constrained to the same system call filters as
58 the parent. If execve(2) is allowed, the existing filters will
59 be preserved across a call to execve(2).
60 In order to use the SECCOMP_SET_MODE_FILTER operation, either
62 the caller must have the CAP_SYS_ADMIN capability in its user
63 namespace, or the thread must already have the no_new_privs bit
64 set. If that bit was not already set by an ancestor of this
65 thread, the thread must make the following call:
66 prctl(PR_SET_NO_NEW_PRIVS, 1);
68 Otherwise, the SECCOMP_SET_MODE_FILTER operation fails and
70 returns EACCES in errno. This requirement ensures that an
71 unprivileged process cannot apply a malicious filter and then
72 invoke a set-user-ID or other privileged program using
73 execve(2), thus potentially compromising that program. (Such a
74 malicious filter might, for example, cause an attempt to use
75 setuid(2) to set the caller's user IDs to nonzero values to
76 instead return 0 without actually making the system call. Thus,
77 the program might be tricked into retaining superuser privileges
78 in circumstances where it is possible to influence it to do dan‐
79 gerous things because it did not actually drop privileges.)
80 If prctl(2) or seccomp() is allowed by the attached filter, fur‐
82 ther filters may be added. This will increase evaluation time,
83 but allows for further reduction of the attack surface during
84 execution of a thread.
85 The SECCOMP_SET_MODE_FILTER operation is available only if the
87 kernel is configured with CONFIG_SECCOMP_FILTER enabled.
88 When flags is 0, this operation is functionally identical to the
90 call:
91 prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, args);
93 The recognized flags are:
95 SECCOMP_FILTER_FLAG_TSYNC
97 When adding a new filter, synchronize all other threads
98 of the calling process to the same seccomp filter tree.
99 A "filter tree" is the ordered list of filters attached
100 to a thread. (Attaching identical filters in separate
101 seccomp() calls results in different filters from this
102 perspective.)
103 If any thread cannot synchronize to the same filter tree,
105 the call will not attach the new seccomp filter, and will
106 fail, returning the first thread ID found that cannot
107 synchronize. Synchronization will fail if another thread
108 in the same process is in SECCOMP_MODE_STRICT or if it
109 has attached new seccomp filters to itself, diverging
110 from the calling thread's filter tree.
111 SECCOMP_FILTER_FLAG_LOG (since Linux 4.14)
113 All filter return actions except SECCOMP_RET_ALLOW should
114 be logged. An administrator may override this filter
115 flag by preventing specific actions from being logged via
116 the /proc/sys/kernel/seccomp/actions_logged file.
117 SECCOMP_GET_ACTION_AVAIL (since Linux 4.14)
119 Test to see if an action is supported by the kernel. This oper‐
120 ation is helpful to confirm that the kernel knows of a more
121 recently added filter return action since the kernel treats all
122 unknown actions as SECCOMP_RET_KILL_PROCESS.
123 The value of flags must be 0, and args must be a pointer to an
125 unsigned 32-bit filter return action.
126 Filters
128 When adding filters via SECCOMP_SET_MODE_FILTER, args points to a fil‐
129 ter program:
130 struct sock_fprog {
132 unsigned short len; /* Number of BPF instructions */
133 struct sock_filter *filter; /* Pointer to array of
134 BPF instructions */
135 };
136 Each program must contain one or more BPF instructions:
138 struct sock_filter { /* Filter block */
140 __u16 code; /* Actual filter code */
141 __u8 jt; /* Jump true */
142 __u8 jf; /* Jump false */
143 __u32 k; /* Generic multiuse field */
144 };
145 When executing the instructions, the BPF program operates on the system
147 call information made available (i.e., use the BPF_ABS addressing mode)
148 as a (read-only) buffer of the following form:
149 struct seccomp_data {
151 int nr; /* System call number */
152 __u32 arch; /* AUDIT_ARCH_* value
153 (see <linux/audit.h>) */
154 __u64 instruction_pointer; /* CPU instruction pointer */
155 __u64 args[6]; /* Up to 6 system call arguments */
156 };
157 Because numbering of system calls varies between architectures and some
159 architectures (e.g., x86-64) allow user-space code to use the calling
160 conventions of multiple architectures, it is usually necessary to ver‐
161 ify the value of the arch field.
162 It is strongly recommended to use a whitelisting approach whenever pos‐
164 sible because such an approach is more robust and simple. A blacklist
165 will have to be updated whenever a potentially dangerous system call is
166 added (or a dangerous flag or option if those are blacklisted), and it
167 is often possible to alter the representation of a value without alter‐
168 ing its meaning, leading to a blacklist bypass. See also Caveats
169 below.
170 The arch field is not unique for all calling conventions. The x86-64
172 ABI and the x32 ABI both use AUDIT_ARCH_X86_64 as arch, and they run on
173 the same processors. Instead, the mask __X32_SYSCALL_BIT is used on
174 the system call number to tell the two ABIs apart.
175 This means that in order to create a seccomp-based blacklist for system
177 calls performed through the x86-64 ABI, it is necessary to not only
178 check that arch equals AUDIT_ARCH_X86_64, but also to explicitly reject
179 all system calls that contain __X32_SYSCALL_BIT in nr.
180 The instruction_pointer field provides the address of the machine-lan‐
182 guage instruction that performed the system call. This might be useful
183 in conjunction with the use of /proc/[pid]/maps to perform checks based
184 on which region (mapping) of the program made the system call. (Proba‐
185 bly, it is wise to lock down the mmap(2) and mprotect(2) system calls
186 to prevent the program from subverting such checks.)
187 When checking values from args against a blacklist, keep in mind that
189 arguments are often silently truncated before being processed, but
190 after the seccomp check. For example, this happens if the i386 ABI is
191 used on an x86-64 kernel: although the kernel will normally not look
192 beyond the 32 lowest bits of the arguments, the values of the full
193 64-bit registers will be present in the seccomp data. A less surpris‐
194 ing example is that if the x86-64 ABI is used to perform a system call
195 that takes an argument of type int, the more-significant half of the
196 argument register is ignored by the system call, but visible in the
197 seccomp data.
198 A seccomp filter returns a 32-bit value consisting of two parts: the
200 most significant 16 bits (corresponding to the mask defined by the con‐
201 stant SECCOMP_RET_ACTION_FULL) contain one of the "action" values
202 listed below; the least significant 16-bits (defined by the constant
203 SECCOMP_RET_DATA) are "data" to be associated with this return value.
204 If multiple filters exist, they are all executed, in reverse order of
206 their addition to the filter tree—that is, the most recently installed
207 filter is executed first. (Note that all filters will be called even
208 if one of the earlier filters returns SECCOMP_RET_KILL. This is done
209 to simplify the kernel code and to provide a tiny speed-up in the exe‐
210 cution of sets of filters by avoiding a check for this uncommon case.)
211 The return value for the evaluation of a given system call is the
212 first-seen action value of highest precedence (along with its accompa‐
213 nying data) returned by execution of all of the filters.
214 In decreasing order of precedence, the action values that may be
216 returned by a seccomp filter are:
217 SECCOMP_RET_KILL_PROCESS (since Linux 4.14)
219 This value results in immediate termination of the process, with
220 a core dump. The system call is not executed. By contrast with
221 SECCOMP_RET_KILL_THREAD below, all threads in the thread group
222 are terminated. (For a discussion of thread groups, see the
223 description of the CLONE_THREAD flag in clone(2).)
224 The process terminates as though killed by a SIGSYS signal.
226 Even if a signal handler has been registered for SIGSYS, the
227 handler will be ignored in this case and the process always ter‐
228 minates. To a parent process that is waiting on this process
229 (using waitpid(2) or similar), the returned wstatus will indi‐
230 cate that its child was terminated as though by a SIGSYS signal.
231 SECCOMP_RET_KILL_THREAD (or SECCOMP_RET_KILL)
233 This value results in immediate termination of the thread that
234 made the system call. The system call is not executed. Other
235 threads in the same thread group will continue to execute.
236 The thread terminates as though killed by a SIGSYS signal. See
238 SECCOMP_RET_KILL_PROCESS above.
239 Before Linux 4.11, any process terminated in this way would not
241 trigger a coredump (even though SIGSYS is documented in sig‐
242 nal(7) as having a default action of termination with a core
243 dump). Since Linux 4.11, a single-threaded process will dump
244 core if terminated in this way.
245 With the addition of SECCOMP_RET_KILL_PROCESS in Linux 4.14,
247 SECCOMP_RET_KILL_THREAD was added as a synonym for SEC‐
248 COMP_RET_KILL, in order to more clearly distinguish the two
249 actions.
250 SECCOMP_RET_TRAP
252 This value results in the kernel sending a thread-directed
253 SIGSYS signal to the triggering thread. (The system call is not
254 executed.) Various fields will be set in the siginfo_t struc‐
255 ture (see sigaction(2)) associated with signal:
256 * si_signo will contain SIGSYS.
258 * si_call_addr will show the address of the system call
260 instruction.
261 * si_syscall and si_arch will indicate which system call was
263 attempted.
264 * si_code will contain SYS_SECCOMP.
266 * si_errno will contain the SECCOMP_RET_DATA portion of the
268 filter return value.
269 The program counter will be as though the system call happened
271 (i.e., the program counter will not point to the system call
272 instruction). The return value register will contain an archi‐
273 tecture-dependent value; if resuming execution, set it to some‐
274 thing appropriate for the system call. (The architecture depen‐
275 dency is because replacing it with ENOSYS could overwrite some
276 useful information.)
277 SECCOMP_RET_ERRNO
279 This value results in the SECCOMP_RET_DATA portion of the fil‐
280 ter's return value being passed to user space as the errno value
281 without executing the system call.
282 SECCOMP_RET_TRACE
284 When returned, this value will cause the kernel to attempt to
285 notify a ptrace(2)-based tracer prior to executing the system
286 call. If there is no tracer present, the system call is not
287 executed and returns a failure status with errno set to ENOSYS.
288 A tracer will be notified if it requests PTRACE_O_TRACESECCOMP
290 using ptrace(PTRACE_SETOPTIONS). The tracer will be notified of
291 a PTRACE_EVENT_SECCOMP and the SECCOMP_RET_DATA portion of the
292 filter's return value will be available to the tracer via
293 PTRACE_GETEVENTMSG.
294 The tracer can skip the system call by changing the system call
296 number to -1. Alternatively, the tracer can change the system
297 call requested by changing the system call to a valid system
298 call number. If the tracer asks to skip the system call, then
299 the system call will appear to return the value that the tracer
300 puts in the return value register.
301 Before kernel 4.8, the seccomp check will not be run again after
303 the tracer is notified. (This means that, on older kernels,
304 seccomp-based sandboxes must not allow use of ptrace(2)—even of
305 other sandboxed processes—without extreme care; ptracers can use
306 this mechanism to escape from the seccomp sandbox.)
307 SECCOMP_RET_LOG (since Linux 4.14)
309 This value results in the system call being executed after the
310 filter return action is logged. An administrator may override
311 the logging of this action via the /proc/sys/kernel/sec‐
312 comp/actions_logged file.
313 SECCOMP_RET_ALLOW
315 This value results in the system call being executed.
316 If an action value other than one of the above is specified, then the
318 filter action is treated as either SECCOMP_RET_KILL_PROCESS (since
319 Linux 4.14) or SECCOMP_RET_KILL_THREAD (in Linux 4.13 and earlier).
320 /proc interfaces
322 The files in the directory /proc/sys/kernel/seccomp provide additional
323 seccomp information and configuration:
324 actions_avail (since Linux 4.14)
326 A read-only ordered list of seccomp filter return actions in
327 string form. The ordering, from left-to-right, is in decreasing
328 order of precedence. The list represents the set of seccomp
329 filter return actions supported by the kernel.
330 actions_logged (since Linux 4.14)
332 A read-write ordered list of seccomp filter return actions that
333 are allowed to be logged. Writes to the file do not need to be
334 in ordered form but reads from the file will be ordered in the
335 same way as the actions_avail file.
336 It is important to note that the value of actions_logged does
338 not prevent certain filter return actions from being logged when
339 the audit subsystem is configured to audit a task. If the
340 action is not found in the actions_logged file, the final deci‐
341 sion on whether to audit the action for that task is ultimately
342 left up to the audit subsystem to decide for all filter return
343 actions other than SECCOMP_RET_ALLOW.
344 The "allow" string is not accepted in the actions_logged file as
346 it is not possible to log SECCOMP_RET_ALLOW actions. Attempting
347 to write "allow" to the file will fail with the error EINVAL.
348 Audit logging of seccomp actions
350 Since Linux 4.14, the kernel provides the facility to log the actions
351 returned by seccomp filters in the audit log. The kernel makes the
352 decision to log an action based on the action type, whether or not the
353 action is present in the actions_logged file, and whether kernel audit‐
354 ing is enabled (e.g., via the kernel boot option audit=1). The rules
355 are as follows:
356 * If the action is SECCOMP_RET_ALLOW, the action is not logged.
358 * Otherwise, if the action is either SECCOMP_RET_KILL_PROCESS or SEC‐
360 COMP_RET_KILL_THREAD, and that action appears in the actions_logged
361 file, the action is logged.
362 * Otherwise, if the filter has requested logging (the SECCOMP_FIL‐
364 TER_FLAG_LOG flag) and the action appears in the actions_logged
365 file, the action is logged.
366 * Otherwise, if kernel auditing is enabled and the process is being
368 audited (autrace(8)), the action is logged.
369 * Otherwise, the action is not logged.
371
373 On success, seccomp() returns 0. On error, if SECCOMP_FIL‐
374 TER_FLAG_TSYNC was used, the return value is the ID of the thread that
375 caused the synchronization failure. (This ID is a kernel thread ID of
376 the type returned by clone(2) and gettid(2).) On other errors, -1 is
377 returned, and errno is set to indicate the cause of the error.
378
380 seccomp() can fail for the following reasons:
381 EACCESS
383 The caller did not have the CAP_SYS_ADMIN capability in its user
384 namespace, or had not set no_new_privs before using SEC‐
385 COMP_SET_MODE_FILTER.
386 EFAULT args was not a valid address.
388 EINVAL operation is unknown or is not supported by this kernel version
390 or configuration.
391 EINVAL The specified flags are invalid for the given operation.
393 EINVAL operation included BPF_ABS, but the specified offset was not
395 aligned to a 32-bit boundary or exceeded sizeof(struct sec‐
396 comp_data).
397 EINVAL A secure computing mode has already been set, and operation dif‐
399 fers from the existing setting.
400 EINVAL operation specified SECCOMP_SET_MODE_FILTER, but the filter pro‐
402 gram pointed to by args was not valid or the length of the fil‐
403 ter program was zero or exceeded BPF_MAXINSNS (4096) instruc‐
404 tions.
405 ENOMEM Out of memory.
407 ENOMEM The total length of all filter programs attached to the calling
409 thread would exceed MAX_INSNS_PER_PATH [22m(32768) instructions.
410 Note that for the purposes of calculating this limit, each
411 already existing filter program incurs an overhead penalty of 4
412 instructions.
413 EOPNOTSUPP
415 operation specified SECCOMP_GET_ACTION_AVAIL, but the kernel
416 does not support the filter return action specified by args.
417 ESRCH Another thread caused a failure during thread sync, but its ID
419 could not be determined.
420
422 The seccomp() system call first appeared in Linux 3.17.
423
425 The seccomp() system call is a nonstandard Linux extension.
426
428 Rather than hand-coding seccomp filters as shown in the example below,
429 you may prefer to employ the libseccomp library, which provides a
430 front-end for generating seccomp filters.
431 The Seccomp field of the /proc/[pid]/status file provides a method of
433 viewing the seccomp mode of a process; see proc(5).
434 seccomp() provides a superset of the functionality provided by the
436 prctl(2) PR_SET_SECCOMP operation (which does not support flags).
437 Since Linux 4.4, the prctl(2) PTRACE_SECCOMP_GET_FILTER operation can
439 be used to dump a process's seccomp filters.
440 Caveats
442 There are various subtleties to consider when applying seccomp filters
443 to a program, including the following:
444 * Some traditional system calls have user-space implementations in the
446 vdso(7) on many architectures. Notable examples include clock_get‐
447 time(2), gettimeofday(2), and time(2). On such architectures, sec‐
448 comp filtering for these system calls will have no effect. (How‐
449 ever, there are cases where the vdso(7) implementations may fall
450 back to invoking the true system call, in which case seccomp filters
451 would see the system call.)
452 * Seccomp filtering is based on system call numbers. However, appli‐
454 cations typically do not directly invoke system calls, but instead
455 call wrapper functions in the C library which in turn invoke the
456 system calls. Consequently, one must be aware of the following:
457 · The glibc wrappers for some traditional system calls may actually
459 employ system calls with different names in the kernel. For
460 example, the exit(2) wrapper function actually employs the
461 exit_group(2) system call, and the fork(2) wrapper function actu‐
462 ally calls clone(2).
463 · The behavior of wrapper functions may vary across architectures,
465 according to the range of system calls provided on those archi‐
466 tectures. In other words, the same wrapper function may invoke
467 different system calls on different architectures.
468 · Finally, the behavior of wrapper functions can change across
470 glibc versions. For example, in older versions, the glibc wrap‐
471 per function for open(2) invoked the system call of the same
472 name, but starting in glibc 2.26, the implementation switched to
473 calling openat(2) on all architectures.
474 The consequence of the above points is that it may be necessary to fil‐
476 ter for a system call other than might be expected. Various manual
477 pages in Section 2 provide helpful details about the differences
478 between wrapper functions and the underlying system calls in subsec‐
479 tions entitled C library/kernel differences.
480 Furthermore, note that the application of seccomp filters even risks
482 causing bugs in an application, when the filters cause unexpected fail‐
483 ures for legitimate operations that the application might need to per‐
484 form. Such bugs may not easily be discovered when testing the seccomp
485 filters if the bugs occur in rarely used application code paths.
486 Seccomp-specific BPF details
488 Note the following BPF details specific to seccomp filters:
489 * The BPF_H and BPF_B size modifiers are not supported: all operations
491 must load and store (4-byte) words (BPF_W).
492 * To access the contents of the seccomp_data buffer, use the BPF_ABS
494 addressing mode modifier.
495 * The BPF_LEN addressing mode modifier yields an immediate mode oper‐
497 and whose value is the size of the seccomp_data buffer.
498
500 The program below accepts four or more arguments. The first three
501 arguments are a system call number, a numeric architecture identifier,
502 and an error number. The program uses these values to construct a BPF
503 filter that is used at run time to perform the following checks:
504 [1] If the program is not running on the specified architecture, the
506 BPF filter causes system calls to fail with the error ENOSYS.
507 [2] If the program attempts to execute the system call with the speci‐
509 fied number, the BPF filter causes the system call to fail, with
510 errno being set to the specified error number.
511 The remaining command-line arguments specify the pathname and addi‐
513 tional arguments of a program that the example program should attempt
514 to execute using execv(3) (a library function that employs the
515 execve(2) system call). Some example runs of the program are shown
516 below.
517 First, we display the architecture that we are running on (x86-64) and
519 then construct a shell function that looks up system call numbers on
520 this architecture:
521 $ uname -m
523 x86_64
524 $ syscall_nr() {
525 cat /usr/src/linux/arch/x86/syscalls/syscall_64.tbl | \
526 awk '$2 != "x32" && $3 == "'$1'" { print $1 }'
527 }
528 When the BPF filter rejects a system call (case [2] above), it causes
530 the system call to fail with the error number specified on the command
531 line. In the experiments shown here, we'll use error number 99:
532 $ errno 99
534 EADDRNOTAVAIL 99 Cannot assign requested address
535 In the following example, we attempt to run the command whoami(1), but
537 the BPF filter rejects the execve(2) system call, so that the command
538 is not even executed:
539 $ syscall_nr execve
541 59
542 $ ./a.out
543 Usage: ./a.out <syscall_nr> <arch> <errno> <prog> [<args>]
544 Hint for <arch>: AUDIT_ARCH_I386: 0x40000003
545 AUDIT_ARCH_X86_64: 0xC000003E
546 $ ./a.out 59 0xC000003E 99 /bin/whoami
547 execv: Cannot assign requested address
548 In the next example, the BPF filter rejects the write(2) system call,
550 so that, although it is successfully started, the whoami(1) command is
551 not able to write output:
552 $ syscall_nr write
554 1
555 $ ./a.out 1 0xC000003E 99 /bin/whoami
556 In the final example, the BPF filter rejects a system call that is not
558 used by the whoami(1) command, so it is able to successfully execute
559 and produce output:
560 $ syscall_nr preadv
562 295
563 $ ./a.out 295 0xC000003E 99 /bin/whoami
564 cecilia
565 Program source
567 #include <errno.h>
568 #include <stddef.h>
569 #include <stdio.h>
570 #include <stdlib.h>
571 #include <unistd.h>
572 #include <linux/audit.h>
573 #include <linux/filter.h>
574 #include <linux/seccomp.h>
575 #include <sys/prctl.h>
576 #define X32_SYSCALL_BIT 0x40000000
578 static int
580 install_filter(int syscall_nr, int t_arch, int f_errno)
581 {
582 unsigned int upper_nr_limit = 0xffffffff;
583 /* Assume that AUDIT_ARCH_X86_64 means the normal x86-64 ABI */
585 if (t_arch == AUDIT_ARCH_X86_64)
586 upper_nr_limit = X32_SYSCALL_BIT - 1;
587 struct sock_filter filter[] = {
589 /* [0] Load architecture from 'seccomp_data' buffer into
590 accumulator */
591 BPF_STMT(BPF_LD | BPF_W | BPF_ABS,
592 (offsetof(struct seccomp_data, arch))),
593 /* [1] Jump forward 5 instructions if architecture does not
595 match 't_arch' */
596 BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, t_arch, 0, 5),
597 /* [2] Load system call number from 'seccomp_data' buffer into
599 accumulator */
600 BPF_STMT(BPF_LD | BPF_W | BPF_ABS,
601 (offsetof(struct seccomp_data, nr))),
602 /* [3] Check ABI - only needed for x86-64 in blacklist use
604 cases. Use BPF_JGT instead of checking against the bit
605 mask to avoid having to reload the syscall number. */
606 BPF_JUMP(BPF_JMP | BPF_JGT | BPF_K, upper_nr_limit, 3, 0),
607 /* [4] Jump forward 1 instruction if system call number
609 does not match 'syscall_nr' */
610 BPF_JUMP(BPF_JMP | BPF_JEQ | BPF_K, syscall_nr, 0, 1),
611 /* [5] Matching architecture and system call: don't execute
613 the system call, and return 'f_errno' in 'errno' */
614 BPF_STMT(BPF_RET | BPF_K,
615 SECCOMP_RET_ERRNO | (f_errno & SECCOMP_RET_DATA)),
616 /* [6] Destination of system call number mismatch: allow other
618 system calls */
619 BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_ALLOW),
620 /* [7] Destination of architecture mismatch: kill task */
622 BPF_STMT(BPF_RET | BPF_K, SECCOMP_RET_KILL),
623 };
624 struct sock_fprog prog = {
626 .len = (unsigned short) (sizeof(filter) / sizeof(filter[0])),
627 .filter = filter,
628 };
629 if (seccomp(SECCOMP_SET_MODE_FILTER, 0, &prog)) {
631 perror("seccomp");
632 return 1;
633 }
634 return 0;
636 }
637 int
639 main(int argc, char **argv)
640 {
641 if (argc < 5) {
642 fprintf(stderr, "Usage: "
643 "%s <syscall_nr> <arch> <errno> <prog> [<args>]\n"
644 "Hint for <arch>: AUDIT_ARCH_I386: 0x%X\n"
645 " AUDIT_ARCH_X86_64: 0x%X\n"
646 "\n", argv[0], AUDIT_ARCH_I386, AUDIT_ARCH_X86_64);
647 exit(EXIT_FAILURE);
648 }
649 if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0)) {
651 perror("prctl");
652 exit(EXIT_FAILURE);
653 }
654 if (install_filter(strtol(argv[1], NULL, 0),
656 strtol(argv[2], NULL, 0),
657 strtol(argv[3], NULL, 0)))
658 exit(EXIT_FAILURE);
659 execv(argv[4], &argv[4]);
661 perror("execv");
662 exit(EXIT_FAILURE);
663 }
664
666 strace(1), bpf(2), prctl(2), ptrace(2), sigaction(2), proc(5), sig‐
667 nal(7), socket(7)
668 Various pages from the libseccomp library, including:
670 scmp_sys_resolver(1), seccomp_init(3), seccomp_load(3), sec‐
671 comp_rule_add(3), and seccomp_export_bpf(3).
672 The kernel source files Documentation/networking/filter.txt and Docu‐
674 mentation/userspace-api/seccomp_filter.rst (or Documentation/prctl/sec‐
675 comp_filter.txt before Linux 4.13).
676 McCanne, S. and Jacobson, V. (1992) The BSD Packet Filter: A New Archi‐
678 tecture for User-level Packet Capture, Proceedings of the USENIX Winter
679 1993 Conference ⟨http://www.tcpdump.org/papers/bpf-usenix93.pdf⟩
680
682 This page is part of release 4.15 of the Linux man-pages project. A
683 description of the project, information about reporting bugs, and the
684 latest version of this page, can be found at
685 https://www.kernel.org/doc/man-pages/.
686Linux 2018-02-02 SECCOMP(2)