1SIGNALFD(2) Linux Programmer's Manual SIGNALFD(2)
2
3
4
6 signalfd - create a file descriptor for accepting signals
7
9 #include <sys/signalfd.h>
10
11 int signalfd(int fd, const sigset_t *mask, int flags);
12
14 signalfd() creates a file descriptor that can be used to accept signals
15 targeted at the caller. This provides an alternative to the use of a
16 signal handler or sigwaitinfo(2), and has the advantage that the file
17 descriptor may be monitored by select(2), poll(2), and epoll(7).
18
19 The mask argument specifies the set of signals that the caller wishes
20 to accept via the file descriptor. This argument is a signal set whose
21 contents can be initialized using the macros described in sigsetops(3).
22 Normally, the set of signals to be received via the file descriptor
23 should be blocked using sigprocmask(2), to prevent the signals being
24 handled according to their default dispositions. It is not possible to
25 receive SIGKILL or SIGSTOP signals via a signalfd file descriptor;
26 these signals are silently ignored if specified in mask.
27
28 If the fd argument is -1, then the call creates a new file descriptor
29 and associates the signal set specified in mask with that file descrip‐
30 tor. If fd is not -1, then it must specify a valid existing signalfd
31 file descriptor, and mask is used to replace the signal set associated
32 with that file descriptor.
33
34 Starting with Linux 2.6.27, the following values may be bitwise ORed in
35 flags to change the behavior of signalfd():
36
37 SFD_NONBLOCK Set the O_NONBLOCK file status flag on the open file
38 description (see open(2)) referred to by the new file
39 descriptor. Using this flag saves extra calls to
40 fcntl(2) to achieve the same result.
41
42 SFD_CLOEXEC Set the close-on-exec (FD_CLOEXEC) flag on the new file
43 descriptor. See the description of the O_CLOEXEC flag in
44 open(2) for reasons why this may be useful.
45
46 In Linux up to version 2.6.26, the flags argument is unused, and must
47 be specified as zero.
48
49 signalfd() returns a file descriptor that supports the following opera‐
50 tions:
51
52 read(2)
53 If one or more of the signals specified in mask is pending for
54 the process, then the buffer supplied to read(2) is used to
55 return one or more signalfd_siginfo structures (see below) that
56 describe the signals. The read(2) returns information for as
57 many signals as are pending and will fit in the supplied buffer.
58 The buffer must be at least sizeof(struct signalfd_siginfo)
59 bytes. The return value of the read(2) is the total number of
60 bytes read.
61
62 As a consequence of the read(2), the signals are consumed, so
63 that they are no longer pending for the process (i.e., will not
64 be caught by signal handlers, and cannot be accepted using sig‐
65 waitinfo(2)).
66
67 If none of the signals in mask is pending for the process, then
68 the read(2) either blocks until one of the signals in mask is
69 generated for the process, or fails with the error EAGAIN if the
70 file descriptor has been made nonblocking.
71
72 poll(2), select(2) (and similar)
73 The file descriptor is readable (the select(2) readfds argument;
74 the poll(2) POLLIN flag) if one or more of the signals in mask
75 is pending for the process.
76
77 The signalfd file descriptor also supports the other file-
78 descriptor multiplexing APIs: pselect(2), ppoll(2), and
79 epoll(7).
80
81 close(2)
82 When the file descriptor is no longer required it should be
83 closed. When all file descriptors associated with the same sig‐
84 nalfd object have been closed, the resources for object are
85 freed by the kernel.
86
87 The signalfd_siginfo structure
88 The format of the signalfd_siginfo structure(s) returned by read(2)s
89 from a signalfd file descriptor is as follows:
90
91 struct signalfd_siginfo {
92 uint32_t ssi_signo; /* Signal number */
93 int32_t ssi_errno; /* Error number (unused) */
94 int32_t ssi_code; /* Signal code */
95 uint32_t ssi_pid; /* PID of sender */
96 uint32_t ssi_uid; /* Real UID of sender */
97 int32_t ssi_fd; /* File descriptor (SIGIO) */
98 uint32_t ssi_tid; /* Kernel timer ID (POSIX timers)
99 uint32_t ssi_band; /* Band event (SIGIO) */
100 uint32_t ssi_overrun; /* POSIX timer overrun count */
101 uint32_t ssi_trapno; /* Trap number that caused signal */
102 int32_t ssi_status; /* Exit status or signal (SIGCHLD) */
103 int32_t ssi_int; /* Integer sent by sigqueue(3) */
104 uint64_t ssi_ptr; /* Pointer sent by sigqueue(3) */
105 uint64_t ssi_utime; /* User CPU time consumed (SIGCHLD) */
106 uint64_t ssi_stime; /* System CPU time consumed
107 (SIGCHLD) */
108 uint64_t ssi_addr; /* Address that generated signal
109 (for hardware-generated signals) */
110 uint16_t ssi_addr_lsb; /* Least significant bit of address
111 (SIGBUS; since Linux 2.6.37)
112 uint8_t pad[X]; /* Pad size to 128 bytes (allow for
113 additional fields in the future) */
114 };
115
116 Each of the fields in this structure is analogous to the similarly
117 named field in the siginfo_t structure. The siginfo_t structure is
118 described in sigaction(2). Not all fields in the returned sig‐
119 nalfd_siginfo structure will be valid for a specific signal; the set of
120 valid fields can be determined from the value returned in the ssi_code
121 field. This field is the analog of the siginfo_t si_code field; see
122 sigaction(2) for details.
123
124 fork(2) semantics
125 After a fork(2), the child inherits a copy of the signalfd file
126 descriptor. A read(2) from the file descriptor in the child will
127 return information about signals queued to the child.
128
129 Semantics of file descriptor passing
130 As with other file descriptors, signalfd file descriptors can be passed
131 to another process via a UNIX domain socket (see unix(7)). In the
132 receiving process, a read(2) from the received file descriptor will
133 return information about signals queued to that process.
134
135 execve(2) semantics
136 Just like any other file descriptor, a signalfd file descriptor remains
137 open across an execve(2), unless it has been marked for close-on-exec
138 (see fcntl(2)). Any signals that were available for reading before the
139 execve(2) remain available to the newly loaded program. (This is anal‐
140 ogous to traditional signal semantics, where a blocked signal that is
141 pending remains pending across an execve(2).)
142
143 Thread semantics
144 The semantics of signalfd file descriptors in a multithreaded program
145 mirror the standard semantics for signals. In other words, when a
146 thread reads from a signalfd file descriptor, it will read the signals
147 that are directed to the thread itself and the signals that are
148 directed to the process (i.e., the entire thread group). (A thread
149 will not be able to read signals that are directed to other threads in
150 the process.)
151
152 epoll(7) semantics
153 If a process adds (via epoll_ctl(2)) a signalfd file descriptor to an
154 epoll(7) instance, then epoll_wait(2) returns events only for signals
155 sent to that process. In particular, if the process then uses fork(2)
156 to create a child process, then the child will be able to read(2) sig‐
157 nals that are sent to it using the signalfd file descriptor, but
158 epoll_wait(2) will not indicate that the signalfd file descriptor is
159 ready. In this scenario, a possible workaround is that after the
160 fork(2), the child process can close the signalfd file descriptor that
161 it inherited from the parent process and then create another signalfd
162 file descriptor and add it to the epoll instance. Alternatively, the
163 parent and the child could delay creating their (separate) signalfd
164 file descriptors and adding them to the epoll instance until after the
165 call to fork(2).
166
168 On success, signalfd() returns a signalfd file descriptor; this is
169 either a new file descriptor (if fd was -1), or fd if fd was a valid
170 signalfd file descriptor. On error, -1 is returned and errno is set to
171 indicate the error.
172
174 EBADF The fd file descriptor is not a valid file descriptor.
175
176 EINVAL fd is not a valid signalfd file descriptor.
177
178 EINVAL flags is invalid; or, in Linux 2.6.26 or earlier, flags is
179 nonzero.
180
181 EMFILE The per-process limit on the number of open file descriptors has
182 been reached.
183
184 ENFILE The system-wide limit on the total number of open files has been
185 reached.
186
187 ENODEV Could not mount (internal) anonymous inode device.
188
189 ENOMEM There was insufficient memory to create a new signalfd file
190 descriptor.
191
193 signalfd() is available on Linux since kernel 2.6.22. Working support
194 is provided in glibc since version 2.8. The signalfd4() system call
195 (see NOTES) is available on Linux since kernel 2.6.27.
196
198 signalfd() and signalfd4() are Linux-specific.
199
201 A process can create multiple signalfd file descriptors. This makes it
202 possible to accept different signals on different file descriptors.
203 (This may be useful if monitoring the file descriptors using select(2),
204 poll(2), or epoll(7): the arrival of different signals will make dif‐
205 ferent file descriptors ready.) If a signal appears in the mask of
206 more than one of the file descriptors, then occurrences of that signal
207 can be read (once) from any one of the file descriptors.
208
209 Attempts to include SIGKILL and SIGSTOP in mask are silently ignored.
210
211 The signal mask employed by a signalfd file descriptor can be viewed
212 via the entry for the corresponding file descriptor in the process's
213 /proc/[pid]/fdinfo directory. See proc(5) for further details.
214
215 Limitations
216 The signalfd mechanism can't be used to receive signals that are syn‐
217 chronously generated, such as the SIGSEGV signal that results from
218 accessing an invalid memory address or the SIGFPE signal that results
219 from an arithmetic error. Such signals can be caught only via signal
220 handler.
221
222 As described above, in normal usage one blocks the signals that will be
223 accepted via signalfd(). If spawning a child process to execute a
224 helper program (that does not need the signalfd file descriptor), then,
225 after the call to fork(2), you will normally want to unblock those sig‐
226 nals before calling execve(2), so that the helper program can see any
227 signals that it expects to see. Be aware, however, that this won't be
228 possible in the case of a helper program spawned behind the scenes by
229 any library function that the program may call. In such cases, one
230 must fall back to using a traditional signal handler that writes to a
231 file descriptor monitored by select(2), poll(2), or epoll(7).
232
233 C library/kernel differences
234 The underlying Linux system call requires an additional argument,
235 size_t sizemask, which specifies the size of the mask argument. The
236 glibc signalfd() wrapper function does not include this argument, since
237 it provides the required value for the underlying system call.
238
239 There are two underlying Linux system calls: signalfd() and the more
240 recent signalfd4(). The former system call does not implement a flags
241 argument. The latter system call implements the flags values described
242 above. Starting with glibc 2.9, the signalfd() wrapper function will
243 use signalfd4() where it is available.
244
246 In kernels before 2.6.25, the ssi_ptr and ssi_int fields are not filled
247 in with the data accompanying a signal sent by sigqueue(3).
248
250 The program below accepts the signals SIGINT and SIGQUIT via a signalfd
251 file descriptor. The program terminates after accepting a SIGQUIT sig‐
252 nal. The following shell session demonstrates the use of the program:
253
254 $ ./signalfd_demo
255 ^C # Control-C generates SIGINT
256 Got SIGINT
257 ^C
258 Got SIGINT
259 ^\ # Control-\ generates SIGQUIT
260 Got SIGQUIT
261 $
262
263 Program source
264
265 #include <sys/signalfd.h>
266 #include <signal.h>
267 #include <unistd.h>
268 #include <stdlib.h>
269 #include <stdio.h>
270
271 #define handle_error(msg) \
272 do { perror(msg); exit(EXIT_FAILURE); } while (0)
273
274 int
275 main(int argc, char *argv[])
276 {
277 sigset_t mask;
278 int sfd;
279 struct signalfd_siginfo fdsi;
280 ssize_t s;
281
282 sigemptyset(&mask);
283 sigaddset(&mask, SIGINT);
284 sigaddset(&mask, SIGQUIT);
285
286 /* Block signals so that they aren't handled
287 according to their default dispositions */
288
289 if (sigprocmask(SIG_BLOCK, &mask, NULL) == -1)
290 handle_error("sigprocmask");
291
292 sfd = signalfd(-1, &mask, 0);
293 if (sfd == -1)
294 handle_error("signalfd");
295
296 for (;;) {
297 s = read(sfd, &fdsi, sizeof(struct signalfd_siginfo));
298 if (s != sizeof(struct signalfd_siginfo))
299 handle_error("read");
300
301 if (fdsi.ssi_signo == SIGINT) {
302 printf("Got SIGINT\n");
303 } else if (fdsi.ssi_signo == SIGQUIT) {
304 printf("Got SIGQUIT\n");
305 exit(EXIT_SUCCESS);
306 } else {
307 printf("Read unexpected signal\n");
308 }
309 }
310 }
311
313 eventfd(2), poll(2), read(2), select(2), sigaction(2), sigprocmask(2),
314 sigwaitinfo(2), timerfd_create(2), sigsetops(3), sigwait(3), epoll(7),
315 signal(7)
316
318 This page is part of release 5.07 of the Linux man-pages project. A
319 description of the project, information about reporting bugs, and the
320 latest version of this page, can be found at
321 https://www.kernel.org/doc/man-pages/.
322
323
324
325Linux 2020-06-09 SIGNALFD(2)