1 ===========================================
2 Seccomp BPF (SECure COMPuting with filters)
3 ===========================================
8 A large number of system calls are exposed to every userland process
9 with many of them going unused for the entire lifetime of the process.
10 As system calls change and mature, bugs are found and eradicated. A
11 certain subset of userland applications benefit by having a reduced set
12 of available system calls. The resulting set reduces the total kernel
13 surface exposed to the application. System call filtering is meant for
14 use with those applications.
16 Seccomp filtering provides a means for a process to specify a filter for
17 incoming system calls. The filter is expressed as a Berkeley Packet
18 Filter (BPF) program, as with socket filters, except that the data
19 operated on is related to the system call being made: system call
20 number and the system call arguments. This allows for expressive
21 filtering of system calls using a filter program language with a long
22 history of being exposed to userland and a straightforward data set.
24 Additionally, BPF makes it impossible for users of seccomp to fall prey
25 to time-of-check-time-of-use (TOCTOU) attacks that are common in system
26 call interposition frameworks. BPF programs may not dereference
27 pointers which constrains all filters to solely evaluating the system
28 call arguments directly.
33 System call filtering isn't a sandbox. It provides a clearly defined
34 mechanism for minimizing the exposed kernel surface. It is meant to be
35 a tool for sandbox developers to use. Beyond that, policy for logical
36 behavior and information flow should be managed with a combination of
37 other system hardening techniques and, potentially, an LSM of your
38 choosing. Expressive, dynamic filters provide further options down this
39 path (avoiding pathological sizes or selecting which of the multiplexed
40 system calls in socketcall() is allowed, for instance) which could be
41 construed, incorrectly, as a more complete sandboxing solution.
46 An additional seccomp mode is added and is enabled using the same
47 prctl(2) call as the strict seccomp. If the architecture has
48 ``CONFIG_HAVE_ARCH_SECCOMP_FILTER``, then filters may be added as below:
51 Now takes an additional argument which specifies a new filter
53 The BPF program will be executed over struct seccomp_data
54 reflecting the system call number, arguments, and other
55 metadata. The BPF program must then return one of the
56 acceptable values to inform the kernel which action should be
61 prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, prog);
63 The 'prog' argument is a pointer to a struct sock_fprog which
64 will contain the filter program. If the program is invalid, the
65 call will return -1 and set errno to ``EINVAL``.
67 If ``fork``/``clone`` and ``execve`` are allowed by @prog, any child
68 processes will be constrained to the same filters and system
69 call ABI as the parent.
71 Prior to use, the task must call ``prctl(PR_SET_NO_NEW_PRIVS, 1)`` or
72 run with ``CAP_SYS_ADMIN`` privileges in its namespace. If these are not
73 true, ``-EACCES`` will be returned. This requirement ensures that filter
74 programs cannot be applied to child processes with greater privileges
75 than the task that installed them.
77 Additionally, if ``prctl(2)`` is allowed by the attached filter,
78 additional filters may be layered on which will increase evaluation
79 time, but allow for further decreasing the attack surface during
80 execution of a process.
82 The above call returns 0 on success and non-zero on error.
87 A seccomp filter may return any of the following values. If multiple
88 filters exist, the return value for the evaluation of a given system
89 call will always use the highest precedent value. (For example,
90 ``SECCOMP_RET_KILL_PROCESS`` will always take precedence.)
92 In precedence order, they are:
94 ``SECCOMP_RET_KILL_PROCESS``:
95 Results in the entire process exiting immediately without executing
96 the system call. The exit status of the task (``status & 0x7f``)
97 will be ``SIGSYS``, not ``SIGKILL``.
99 ``SECCOMP_RET_KILL_THREAD``:
100 Results in the task exiting immediately without executing the
101 system call. The exit status of the task (``status & 0x7f``) will
102 be ``SIGSYS``, not ``SIGKILL``.
104 ``SECCOMP_RET_TRAP``:
105 Results in the kernel sending a ``SIGSYS`` signal to the triggering
106 task without executing the system call. ``siginfo->si_call_addr``
107 will show the address of the system call instruction, and
108 ``siginfo->si_syscall`` and ``siginfo->si_arch`` will indicate which
109 syscall was attempted. The program counter will be as though
110 the syscall happened (i.e. it will not point to the syscall
111 instruction). The return value register will contain an arch-
112 dependent value -- if resuming execution, set it to something
113 sensible. (The architecture dependency is because replacing
114 it with ``-ENOSYS`` could overwrite some useful information.)
116 The ``SECCOMP_RET_DATA`` portion of the return value will be passed
119 ``SIGSYS`` triggered by seccomp will have a si_code of ``SYS_SECCOMP``.
121 ``SECCOMP_RET_ERRNO``:
122 Results in the lower 16-bits of the return value being passed
123 to userland as the errno without executing the system call.
125 ``SECCOMP_RET_TRACE``:
126 When returned, this value will cause the kernel to attempt to
127 notify a ``ptrace()``-based tracer prior to executing the system
128 call. If there is no tracer present, ``-ENOSYS`` is returned to
129 userland and the system call is not executed.
131 A tracer will be notified if it requests ``PTRACE_O_TRACESECCOM``P
132 using ``ptrace(PTRACE_SETOPTIONS)``. The tracer will be notified
133 of a ``PTRACE_EVENT_SECCOMP`` and the ``SECCOMP_RET_DATA`` portion of
134 the BPF program return value will be available to the tracer
135 via ``PTRACE_GETEVENTMSG``.
137 The tracer can skip the system call by changing the syscall number
138 to -1. Alternatively, the tracer can change the system call
139 requested by changing the system call to a valid syscall number. If
140 the tracer asks to skip the system call, then the system call will
141 appear to return the value that the tracer puts in the return value
144 The seccomp check will not be run again after the tracer is
145 notified. (This means that seccomp-based sandboxes MUST NOT
146 allow use of ptrace, even of other sandboxed processes, without
147 extreme care; ptracers can use this mechanism to escape.)
150 Results in the system call being executed after it is logged. This
151 should be used by application developers to learn which syscalls their
152 application needs without having to iterate through multiple test and
153 development cycles to build the list.
155 This action will only be logged if "log" is present in the
156 actions_logged sysctl string.
158 ``SECCOMP_RET_ALLOW``:
159 Results in the system call being executed.
161 If multiple filters exist, the return value for the evaluation of a
162 given system call will always use the highest precedent value.
164 Precedence is only determined using the ``SECCOMP_RET_ACTION`` mask. When
165 multiple filters return values of the same precedence, only the
166 ``SECCOMP_RET_DATA`` from the most recently installed filter will be
172 The biggest pitfall to avoid during use is filtering on system call
173 number without checking the architecture value. Why? On any
174 architecture that supports multiple system call invocation conventions,
175 the system call numbers may vary based on the specific invocation. If
176 the numbers in the different calling conventions overlap, then checks in
177 the filters may be abused. Always check the arch value!
182 The ``samples/seccomp/`` directory contains both an x86-specific example
183 and a more generic example of a higher level macro interface for BPF
189 Seccomp's sysctl files can be found in the ``/proc/sys/kernel/seccomp/``
190 directory. Here's a description of each file in that directory:
193 A read-only ordered list of seccomp return values (refer to the
194 ``SECCOMP_RET_*`` macros above) in string form. The ordering, from
195 left-to-right, is the least permissive return value to the most
196 permissive return value.
198 The list represents the set of seccomp return values supported
199 by the kernel. A userspace program may use this list to
200 determine if the actions found in the ``seccomp.h``, when the
201 program was built, differs from the set of actions actually
202 supported in the current running kernel.
205 A read-write ordered list of seccomp return values (refer to the
206 ``SECCOMP_RET_*`` macros above) that are allowed to be logged. Writes
207 to the file do not need to be in ordered form but reads from the file
208 will be ordered in the same way as the actions_avail sysctl.
210 It is important to note that the value of ``actions_logged`` does not
211 prevent certain actions from being logged when the audit subsystem is
212 configured to audit a task. If the action is not found in
213 ``actions_logged`` list, the final decision on whether to audit the
214 action for that task is ultimately left up to the audit subsystem to
215 decide for all seccomp return values other than ``SECCOMP_RET_ALLOW``.
217 The ``allow`` string is not accepted in the ``actions_logged`` sysctl
218 as it is not possible to log ``SECCOMP_RET_ALLOW`` actions. Attempting
219 to write ``allow`` to the sysctl will result in an EINVAL being
222 Adding architecture support
223 ===========================
225 See ``arch/Kconfig`` for the authoritative requirements. In general, if an
226 architecture supports both ptrace_event and seccomp, it will be able to
227 support seccomp filter with minor fixup: ``SIGSYS`` support and seccomp return
228 value checking. Then it must just add ``CONFIG_HAVE_ARCH_SECCOMP_FILTER``
229 to its arch-specific Kconfig.
236 The vDSO can cause some system calls to run entirely in userspace,
237 leading to surprises when you run programs on different machines that
238 fall back to real syscalls. To minimize these surprises on x86, make
240 ``/sys/devices/system/clocksource/clocksource0/current_clocksource`` set to
241 something like ``acpi_pm``.
243 On x86-64, vsyscall emulation is enabled by default. (vsyscalls are
244 legacy variants on vDSO calls.) Currently, emulated vsyscalls will
245 honor seccomp, with a few oddities:
247 - A return value of ``SECCOMP_RET_TRAP`` will set a ``si_call_addr`` pointing to
248 the vsyscall entry for the given call and not the address after the
249 'syscall' instruction. Any code which wants to restart the call
250 should be aware that (a) a ret instruction has been emulated and (b)
251 trying to resume the syscall will again trigger the standard vsyscall
252 emulation security checks, making resuming the syscall mostly
255 - A return value of ``SECCOMP_RET_TRACE`` will signal the tracer as usual,
256 but the syscall may not be changed to another system call using the
257 orig_rax register. It may only be changed to -1 order to skip the
258 currently emulated call. Any other change MAY terminate the process.
259 The rip value seen by the tracer will be the syscall entry address;
260 this is different from normal behavior. The tracer MUST NOT modify
261 rip or rsp. (Do not rely on other changes terminating the process.
262 They might work. For example, on some kernels, choosing a syscall
263 that only exists in future kernels will be correctly emulated (by
264 returning ``-ENOSYS``).
266 To detect this quirky behavior, check for ``addr & ~0x0C00 ==
267 0xFFFFFFFFFF600000``. (For ``SECCOMP_RET_TRACE``, use rip. For
268 ``SECCOMP_RET_TRAP``, use ``siginfo->si_call_addr``.) Do not check any other
269 condition: future kernels may improve vsyscall emulation and current
270 kernels in vsyscall=native mode will behave differently, but the
271 instructions at ``0xF...F600{0,4,8,C}00`` will not be system calls in these
274 Note that modern systems are unlikely to use vsyscalls at all -- they
275 are a legacy feature and they are considerably slower than standard
276 syscalls. New code will use the vDSO, and vDSO-issued system calls
277 are indistinguishable from normal system calls.