1SIGACTION(2) Linux Programmer's Manual SIGACTION(2)
2
6 sigaction - examine and change a signal action
7
9 #include <signal.h>
10 int sigaction(int signum, const struct sigaction *act,
12 struct sigaction *oldact);
13 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
15 sigaction(): _POSIX_C_SOURCE >= 1 || _XOPEN_SOURCE || _POSIX_SOURCE
17 siginfo_t: _POSIX_C_SOURCE >= 199309L
19
21 The sigaction() system call is used to change the action taken by a
22 process on receipt of a specific signal. (See signal(7) for an over‐
23 view of signals.)
24 signum specifies the signal and can be any valid signal except SIGKILL
26 and SIGSTOP.
27 If act is non-NULL, the new action for signal signum is installed from
29 act. If oldact is non-NULL, the previous action is saved in oldact.
30 The sigaction structure is defined as something like:
32 struct sigaction {
34 void (*sa_handler)(int);
35 void (*sa_sigaction)(int, siginfo_t *, void *);
36 sigset_t sa_mask;
37 int sa_flags;
38 void (*sa_restorer)(void);
39 };
40 On some architectures a union is involved: do not assign to both
42 sa_handler and sa_sigaction.
43 The sa_restorer element is obsolete and should not be used. POSIX does
45 not specify a sa_restorer element.
46 sa_handler specifies the action to be associated with signum and may be
48 SIG_DFL for the default action, SIG_IGN to ignore this signal, or a
49 pointer to a signal handling function. This function receives the sig‐
50 nal number as its only argument.
51 If SA_SIGINFO is specified in sa_flags, then sa_sigaction (instead of
53 sa_handler) specifies the signal-handling function for signum. This
54 function receives the signal number as its first argument, a pointer to
55 a siginfo_t as its second argument and a pointer to a ucontext_t (cast
56 to void *) as its third argument. (Commonly, the handler function
57 doesn't make any use of the third argument. See getcontext(3) for fur‐
58 ther information about ucontext_t.)
59 sa_mask specifies a mask of signals which should be blocked (i.e.,
61 added to the signal mask of the thread in which the signal handler is
62 invoked) during execution of the signal handler. In addition, the sig‐
63 nal which triggered the handler will be blocked, unless the SA_NODEFER
64 flag is used.
65 sa_flags specifies a set of flags which modify the behavior of the sig‐
67 nal. It is formed by the bitwise OR of zero or more of the following:
68 SA_NOCLDSTOP
70 If signum is SIGCHLD, do not receive notification when child
71 processes stop (i.e., when they receive one of SIGSTOP,
72 SIGTSTP, SIGTTIN or SIGTTOU) or resume (i.e., they receive
73 SIGCONT) (see wait(2)). This flag is meaningful only when
74 establishing a handler for SIGCHLD.
75 SA_NOCLDWAIT (since Linux 2.6)
77 If signum is SIGCHLD, do not transform children into zombies
78 when they terminate. See also waitpid(2). This flag is
79 meaningful only when establishing a handler for SIGCHLD, or
80 when setting that signal's disposition to SIG_DFL.
81 If the SA_NOCLDWAIT flag is set when establishing a handler
83 for SIGCHLD, POSIX.1 leaves it unspecified whether a SIGCHLD
84 signal is generated when a child process terminates. On
85 Linux, a SIGCHLD signal is generated in this case; on some
86 other implementations, it is not.
87 SA_NODEFER
89 Do not prevent the signal from being received from within
90 its own signal handler. This flag is meaningful only when
91 establishing a signal handler. SA_NOMASK is an obsolete,
92 nonstandard synonym for this flag.
93 SA_ONSTACK
95 Call the signal handler on an alternate signal stack pro‐
96 vided by sigaltstack(2). If an alternate stack is not
97 available, the default stack will be used. This flag is
98 meaningful only when establishing a signal handler.
99 SA_RESETHAND
101 Restore the signal action to the default upon entry to the
102 signal handler. This flag is meaningful only when estab‐
103 lishing a signal handler. SA_ONESHOT is an obsolete, non‐
104 standard synonym for this flag.
105 SA_RESTART
107 Provide behavior compatible with BSD signal semantics by
108 making certain system calls restartable across signals.
109 This flag is meaningful only when establishing a signal han‐
110 dler. See signal(7) for a discussion of system call
111 restarting.
112 SA_SIGINFO (since Linux 2.2)
114 The signal handler takes three arguments, not one. In this
115 case, sa_sigaction should be set instead of sa_handler.
116 This flag is meaningful only when establishing a signal han‐
117 dler.
118 The siginfo_t argument to sa_sigaction is a struct with the following
120 elements:
121 siginfo_t {
123 int si_signo; /* Signal number */
124 int si_errno; /* An errno value */
125 int si_code; /* Signal code */
126 int si_trapno; /* Trap number that caused
127 hardware-generated signal
128 (unused on most architectures) */
129 pid_t si_pid; /* Sending process ID */
130 uid_t si_uid; /* Real user ID of sending process */
131 int si_status; /* Exit value or signal */
132 clock_t si_utime; /* User time consumed */
133 clock_t si_stime; /* System time consumed */
134 sigval_t si_value; /* Signal value */
135 int si_int; /* POSIX.1b signal */
136 void *si_ptr; /* POSIX.1b signal */
137 int si_overrun; /* Timer overrun count; POSIX.1b timers */
138 int si_timerid; /* Timer ID; POSIX.1b timers */
139 void *si_addr; /* Memory location which caused fault */
140 long si_band; /* Band event (was int in
141 glibc 2.3.2 and earlier) */
142 int si_fd; /* File descriptor */
143 short si_addr_lsb; /* Least significant bit of address
144 (since Linux 2.6.32) */
145 }
146 si_signo, si_errno and si_code are defined for all signals. (si_errno
148 is generally unused on Linux.) The rest of the struct may be a union,
149 so that one should read only the fields that are meaningful for the
150 given signal:
151 * Signals sent with kill(2) and sigqueue(3) fill in si_pid and si_uid.
153 In addition, signals sent with sigqueue(3) fill in si_int and si_ptr
154 with the values specified by the sender of the signal; see
155 sigqueue(3) for more details.
156 * Signals sent by POSIX.1b timers (since Linux 2.6) fill in si_overrun
158 and si_timerid. The si_timerid field is an internal ID used by the
159 kernel to identify the timer; it is not the same as the timer ID
160 returned by timer_create(2). The si_overrun field is the timer over‐
161 run count; this is the same information as is obtained by a call to
162 timer_getoverrun(2). These fields are nonstandard Linux extensions.
163 * Signals sent for message queue notification (see the description of
165 SIGEV_SIGNAL in mq_notify(3)) fill in si_int/si_ptr, with the
166 sigev_value supplied to mq_notify(3); si_pid, with the process ID of
167 the message sender; and si_uid, with the real user ID of the message
168 sender.
169 * SIGCHLD fills in si_pid, si_uid, si_status, si_utime and si_stime,
171 providing information about the child. The si_pid field is the
172 process ID of the child; si_uid is the child's real user ID. The
173 si_status field contains the exit status of the child (if si_code is
174 CLD_EXITED), or the signal number that caused the process to change
175 state. The si_utime and si_stime contain the user and system CPU
176 time used by the child process; these fields do not include the times
177 used by waited-for children (unlike getrusage(2) and time(2)). In
178 kernels up to 2.6, and since 2.6.27, these fields report CPU time in
179 units of sysconf(_SC_CLK_TCK). In 2.6 kernels before 2.6.27, a bug
180 meant that these fields reported time in units of the (configurable)
181 system jiffy (see time(7)).
182 * SIGILL, SIGFPE, SIGSEGV, SIGBUS, and SIGTRAP fill in si_addr with the
184 address of the fault. On some architectures, these signals also fill
185 in the si_trapno filed. Some suberrors of SIGBUS, in particular
186 BUS_MCEERR_AO and BUS_MCEERR_AR, also fill in si_addr_lsb. This
187 field indicates the least significant bit of the reported address and
188 therefore the extent of the corruption. For example, if a full page
189 was corrupted, si_addr_lsb contains log2(sysconf(_SC_PAGESIZE)).
190 BUS_MCERR_* and si_addr_lsb are Linux-specific extensions.
191 * SIGIO/SIGPOLL (the two names are synonyms on Linux) fills in si_band
193 and si_fd. The si_band event is a bit mask containing the same val‐
194 ues as are filled in the revents field by poll(2). The si_fd field
195 indicates the file descriptor for which the I/O event occurred.
196 si_code is a value (not a bit mask) indicating why this signal was
198 sent. The following list shows the values which can be placed in
199 si_code for any signal, along with reason that the signal was gener‐
200 ated.
201 SI_USER kill(2)
203 SI_KERNEL Sent by the kernel.
205 SI_QUEUE sigqueue(3)
207 SI_TIMER POSIX timer expired
209 SI_MESGQ POSIX message queue state changed (since Linux
211 2.6.6); see mq_notify(3)
212 SI_ASYNCIO AIO completed
214 SI_SIGIO Queued SIGIO (only in kernels up to Linux 2.2; from
216 Linux 2.4 onward SIGIO/SIGPOLL fills in si_code as
217 described below).
218 SI_TKILL tkill(2) or tgkill(2) (since Linux 2.4.19)
220 The following values can be placed in si_code for a SIGILL signal:
222 ILL_ILLOPC illegal opcode
224 ILL_ILLOPN illegal operand
226 ILL_ILLADR illegal addressing mode
228 ILL_ILLTRP illegal trap
230 ILL_PRVOPC privileged opcode
232 ILL_PRVREG privileged register
234 ILL_COPROC coprocessor error
236 ILL_BADSTK internal stack error
238 The following values can be placed in si_code for a SIGFPE signal:
240 FPE_INTDIV integer divide by zero
242 FPE_INTOVF integer overflow
244 FPE_FLTDIV floating-point divide by zero
246 FPE_FLTOVF floating-point overflow
248 FPE_FLTUND floating-point underflow
250 FPE_FLTRES floating-point inexact result
252 FPE_FLTINV floating-point invalid operation
254 FPE_FLTSUB subscript out of range
256 The following values can be placed in si_code for a SIGSEGV signal:
258 SEGV_MAPERR address not mapped to object
260 SEGV_ACCERR invalid permissions for mapped object
262 The following values can be placed in si_code for a SIGBUS signal:
264 BUS_ADRALN invalid address alignment
266 BUS_ADRERR nonexistent physical address
268 BUS_OBJERR object-specific hardware error
270 BUS_MCEERR_AR (since Linux 2.6.32)
272 Hardware memory error consumed on a machine check;
273 action required.
274 BUS_MCEERR_AO (since Linux 2.6.32)
276 Hardware memory error detected in process but not
277 consumed; action optional.
278 The following values can be placed in si_code for a SIGTRAP signal:
280 TRAP_BRKPT process breakpoint
282 TRAP_TRACE process trace trap
284 TRAP_BRANCH (since Linux 2.4)
286 process taken branch trap
287 TRAP_HWBKPT (since Linux 2.4)
289 hardware breakpoint/watchpoint
290 The following values can be placed in si_code for a SIGCHLD signal:
292 CLD_EXITED child has exited
294 CLD_KILLED child was killed
296 CLD_DUMPED child terminated abnormally
298 CLD_TRAPPED traced child has trapped
300 CLD_STOPPED child has stopped
302 CLD_CONTINUED stopped child has continued (since Linux 2.6.9)
304 The following values can be placed in si_code for a SIGIO/SIGPOLL sig‐
306 nal:
307 POLL_IN data input available
309 POLL_OUT output buffers available
311 POLL_MSG input message available
313 POLL_ERR I/O error
315 POLL_PRI high priority input available
317 POLL_HUP device disconnected
319
321 sigaction() returns 0 on success; on error, -1 is returned, and errno
322 is set to indicate the error.
323
325 EFAULT act or oldact points to memory which is not a valid part of the
326 process address space.
327 EINVAL An invalid signal was specified. This will also be generated if
329 an attempt is made to change the action for SIGKILL or SIGSTOP,
330 which cannot be caught or ignored.
331
333 POSIX.1-2001, SVr4.
334
336 A child created via fork(2) inherits a copy of its parent's signal dis‐
337 positions. During an execve(2), the dispositions of handled signals
338 are reset to the default; the dispositions of ignored signals are left
339 unchanged.
340 According to POSIX, the behavior of a process is undefined after it
342 ignores a SIGFPE, SIGILL, or SIGSEGV signal that was not generated by
343 kill(2) or raise(3). Integer division by zero has undefined result.
344 On some architectures it will generate a SIGFPE signal. (Also dividing
345 the most negative integer by -1 may generate SIGFPE.) Ignoring this
346 signal might lead to an endless loop.
347 POSIX.1-1990 disallowed setting the action for SIGCHLD to SIG_IGN.
349 POSIX.1-2001 allows this possibility, so that ignoring SIGCHLD can be
350 used to prevent the creation of zombies (see wait(2)). Nevertheless,
351 the historical BSD and System V behaviors for ignoring SIGCHLD differ,
352 so that the only completely portable method of ensuring that terminated
353 children do not become zombies is to catch the SIGCHLD signal and per‐
354 form a wait(2) or similar.
355 POSIX.1-1990 specified only SA_NOCLDSTOP. POSIX.1-2001 added SA_NOCLD‐
357 WAIT, SA_RESETHAND, SA_NODEFER, and SA_SIGINFO. Use of these latter
358 values in sa_flags may be less portable in applications intended for
359 older UNIX implementations.
360 The SA_RESETHAND flag is compatible with the SVr4 flag of the same
362 name.
363 The SA_NODEFER flag is compatible with the SVr4 flag of the same name
365 under kernels 1.3.9 and newer. On older kernels the Linux implementa‐
366 tion allowed the receipt of any signal, not just the one we are
367 installing (effectively overriding any sa_mask settings).
368 sigaction() can be called with a NULL second argument to query the cur‐
370 rent signal handler. It can also be used to check whether a given sig‐
371 nal is valid for the current machine by calling it with NULL second and
372 third arguments.
373 It is not possible to block SIGKILL or SIGSTOP (by specifying them in
375 sa_mask). Attempts to do so are silently ignored.
376 See sigsetops(3) for details on manipulating signal sets.
378 See signal(7) for a list of the async-signal-safe functions that can be
380 safely called inside from inside a signal handler.
381 Undocumented
383 Before the introduction of SA_SIGINFO it was also possible to get some
384 additional information, namely by using a sa_handler with second argu‐
385 ment of type struct sigcontext. See the relevant Linux kernel sources
386 for details. This use is obsolete now.
387
389 In kernels up to and including 2.6.13, specifying SA_NODEFER in
390 sa_flags prevents not only the delivered signal from being masked dur‐
391 ing execution of the handler, but also the signals specified in
392 sa_mask. This bug was fixed in kernel 2.6.14.
393
395 See mprotect(2).
396
398 kill(1), kill(2), killpg(2), pause(2), restart_syscall(2), sigalt‐
399 stack(2), signal(2), signalfd(2), sigpending(2), sigprocmask(2), sig‐
400 suspend(2), wait(2), raise(3), siginterrupt(3), sigqueue(3), sigse‐
401 tops(3), sigvec(3), core(5), signal(7)
402
404 This page is part of release 3.53 of the Linux man-pages project. A
405 description of the project, and information about reporting bugs, can
406 be found at http://www.kernel.org/doc/man-pages/.
407Linux 2013-07-30 SIGACTION(2)