2 .\" Copyright (c) 1994,1995 Mike Battersby <mib@deakin.edu.au>
3 .\" and Copyright 2004, 2005 Michael Kerrisk <mtk.manpages@gmail.com>
4 .\" based on work by faith@cs.unc.edu
6 .\" Permission is granted to make and distribute verbatim copies of this
7 .\" manual provided the copyright notice and this permission notice are
8 .\" preserved on all copies.
10 .\" Permission is granted to copy and distribute modified versions of this
11 .\" manual under the conditions for verbatim copying, provided that the
12 .\" entire resulting derived work is distributed under the terms of a
13 .\" permission notice identical to this one.
15 .\" Since the Linux kernel and libraries are constantly changing, this
16 .\" manual page may be incorrect or out-of-date. The author(s) assume no
17 .\" responsibility for errors or omissions, or for damages resulting from
18 .\" the use of the information contained herein. The author(s) may not
19 .\" have taken the same level of care in the production of this manual,
20 .\" which is licensed free of charge, as they might when working
23 .\" Formatted or processed versions of this manual, if unaccompanied by
24 .\" the source, must acknowledge the copyright and authors of this work.
26 .\" Modified, aeb, 960424
27 .\" Modified Fri Jan 31 17:31:20 1997 by Eric S. Raymond <esr@thyrsus.com>
28 .\" Modified Thu Nov 26 02:12:45 1998 by aeb - add SIGCHLD stuff.
29 .\" Modified Sat May 8 17:40:19 1999 by Matthew Wilcox
30 .\" add POSIX.1b signals
31 .\" Modified Sat Dec 29 01:44:52 2001 by Evan Jones <ejones@uwaterloo.ca>
33 .\" Modified 2004-11-11 by Michael Kerrisk <mtk.manpages@gmail.com>
34 .\" Added mention of SIGCONT under SA_NOCLDSTOP
35 .\" Added SA_NOCLDWAIT
36 .\" Modified 2004-11-17 by Michael Kerrisk <mtk.manpages@gmail.com>
37 .\" Updated discussion for POSIX.1-2001 and SIGCHLD and sa_flags.
39 .\" 2004-12-09, mtk, added SI_TKILL + other minor changes
40 .\" 2005-09-15, mtk, split sigpending(), sigprocmask(), sigsuspend()
41 .\" out of this page into separate pages.
43 .TH SIGACTION 2 2009-07-25 "Linux" "Linux Programmer's Manual"
45 sigaction \- examine and change a signal action
48 .B #include <signal.h>
50 .BI "int sigaction(int " signum ", const struct sigaction *" act ,
51 .BI " struct sigaction *" oldact );
55 Feature Test Macro Requirements for glibc (see
56 .BR feature_test_macros (7)):
61 _POSIX_C_SOURCE\ >=\ 1 || _XOPEN_SOURCE || _POSIX_SOURCE
66 system call is used to change the action taken by a process on
67 receipt of a specific signal.
70 for an overview of signals.)
73 specifies the signal and can be any valid signal except
80 is non-null, the new action for signal
86 is non-null, the previous action is saved in
91 structure is defined as something like:
96 void (*sa_handler)(int);
97 void (*sa_sigaction)(int, siginfo_t *, void *);
100 void (*sa_restorer)(void);
105 On some architectures a union is involved: do not assign to both
112 element is obsolete and should not be used.
113 POSIX does not specify a
118 specifies the action to be associated with
122 for the default action,
124 to ignore this signal, or a pointer to a signal handling function.
125 This function receives the signal number as its only argument.
135 specifies the signal-handling function for
137 This function receives the signal number as its first argument, a
140 as its second argument and a pointer to a
142 (cast to \fIvoid\ *\fP) as its third argument.
145 specifies a mask of signals which should be blocked
146 (i.e., added to the signal mask of the thread in which
147 the signal handler is invoked)
148 during execution of the signal handler.
149 In addition, the signal which triggered the handler
150 will be blocked, unless the
155 specifies a set of flags which modify the behavior of the signal.
156 It is formed by the bitwise OR of zero or more of the following:
164 do not receive notification when child processes stop (i.e., when they
166 .BR SIGSTOP ", " SIGTSTP ", " SIGTTIN
169 or resume (i.e., they receive
173 This flag is only meaningful when establishing a handler for
176 .BR SA_NOCLDWAIT " (Since Linux 2.6)"
177 .\" To be precise: Linux 2.5.60 -- MTK
182 do not transform children into zombies when they terminate.
185 This flag is only meaningful when establishing a handler for
187 or when setting that signal's disposition to
192 flag is set when establishing a handler for
194 POSIX.1 leaves it unspecified whether a
196 signal is generated when a child process terminates.
199 signal is generated in this case;
200 on some other implementations, it is not.
203 Do not prevent the signal from being received from within its own signal
205 This flag is only meaningful when establishing a signal handler.
207 is an obsolete, nonstandard synonym for this flag.
210 Call the signal handler on an alternate signal stack provided by
212 If an alternate stack is not available, the default stack will be used.
213 This flag is only meaningful when establishing a signal handler.
216 Restore the signal action to the default state once the signal handler
218 This flag is only meaningful when establishing a signal handler.
220 is an obsolete, nonstandard synonym for this flag.
223 Provide behavior compatible with BSD signal semantics by making certain
224 system calls restartable across signals.
225 This flag is only meaningful when establishing a signal handler.
228 for a discussion of system call restarting.
230 .BR SA_SIGINFO " (since Linux 2.2)"
231 The signal handler takes 3 arguments, not one.
234 should be set instead of
236 This flag is only meaningful when establishing a signal handler.
239 .\" field was added in Linux 2.1.86.)
246 is a struct with the following elements:
251 int si_signo; /* Signal number */
252 int si_errno; /* An errno value */
253 int si_code; /* Signal code */
254 int si_trapno; /* Trap number that caused
255 hardware-generated signal
256 (unused on most architectures) */
258 .\" si_trapno seems to be only used on SPARC and Alpha;
259 .\" this page could use a little more detail on its purpose there.
260 pid_t si_pid; /* Sending process ID */
261 uid_t si_uid; /* Real user ID of sending process */
262 int si_status; /* Exit value or signal */
263 clock_t si_utime; /* User time consumed */
264 clock_t si_stime; /* System time consumed */
265 sigval_t si_value; /* Signal value */
266 int si_int; /* POSIX.1b signal */
267 void *si_ptr; /* POSIX.1b signal */
268 int si_overrun; /* Timer overrun count; POSIX.1b timers */
269 int si_timerid; /* Timer ID; POSIX.1b timers */
270 .\" In the kernel: si_tid
271 void *si_addr; /* Memory location which caused fault */
272 int si_band; /* Band event */
273 int si_fd; /* File descriptor */
278 .IR si_signo ", " si_errno " and " si_code
279 are defined for all signals.
281 is generally unused on Linux.)
282 The rest of the struct may be a union, so that one should only
283 read the fields that are meaningful for the given signal:
288 .IR si_pid " and " si_uid .
290 POSIX.1b timers (since Linux 2.6) fill in
296 field is an internal ID used by the kernel to identify
297 the timer; it is not the same as the timer ID returned by
298 .BR timer_create (2).
302 .IR si_status ", " si_utime " and " si_stime .
307 fields do not include the times used by waited-for children (unlike
311 In kernels up to 2.6, and since 2.6.27, these fields report
313 .IR sysconf(_SC_CLK_TCK) .
314 In 2.6 kernels before 2.6.27,
315 a bug meant that these fields reported time in units
316 of the (configurable) system jiffy (see
319 .\" When si_utime and si_stime where originally implemented, the
320 .\" measurement unit was HZ, which was the same as clock ticks
321 .\" (sysconf(_SC_CLK_TCK)). In 2.6, HZ became configurable, and
322 .\" was *still* used as the unit to return the info these fields,
323 .\" with the result that the field values depended on the the
324 .\" configured HZ. Of course, the should have been measured in
325 .\" USER_HZ instead, so that sysconf(_SC_CLK_TCK) could be used to
326 .\" convert to seconds. I have a queued patch to fix this:
327 .\" http://thread.gmane.org/gmane.linux.kernel/698061/ .
328 .\" This patch made it into 2.6.27.
329 .\" But note that these fields still don't return the times of
330 .\" waited-for children (as is done by getrusage() and times()
331 .\" and wait4()). Solaris 8 does include child times.
333 .IR si_int " and " si_ptr
334 are specified by the sender of the POSIX.1b signal.
346 with the address of the fault.
349 .IR si_band " and " si_fd .
352 is a value (not a bit mask)
353 indicating why this signal was sent.
354 The following list shows the values which can be placed in
356 for any signal, along with reason that the signal was generated.
374 POSIX message queue state changed (since Linux 2.6.6); see
388 .\" SI_DETHREAD is defined in 2.6.9 sources, but isn't implemented
389 .\" It appears to have been an idea that was tried during 2.5.6
390 .\" through to 2.5.24 and then was backed out.
393 The following values can be placed in
407 illegal addressing mode
425 The following values can be placed in
433 integer divide by zero
439 floating-point divide by zero
442 floating-point overflow
445 floating-point underflow
448 floating-point inexact result
451 floating-point invalid operation
454 subscript out of range
457 The following values can be placed in
465 address not mapped to object
468 invalid permissions for mapped object
471 The following values can be placed in
479 invalid address alignment
482 nonexistent physical address
485 object-specific hardware error
488 The following values can be placed in
502 The following values can be placed in
516 child terminated abnormally
519 traced child has trapped
525 stopped child has continued (since Linux 2.6.9)
528 The following values can be placed in
539 output buffers available
542 input message available
548 high priority input available
555 returns 0 on success and \-1 on error.
559 .IR act " or " oldact
560 points to memory which is not a valid part of the process address space.
563 An invalid signal was specified.
564 This will also be generated if an attempt
565 is made to change the action for
566 .BR SIGKILL " or " SIGSTOP ", "
567 which cannot be caught or ignored.
570 .\" SVr4 does not document the EINTR condition.
574 inherits a copy of its parent's signal dispositions.
577 the dispositions of handled signals are reset to the default;
578 the dispositions of ignored signals are left unchanged.
580 According to POSIX, the behavior of a process is undefined after it
586 signal that was not generated by
590 Integer division by zero has undefined result.
591 On some architectures it will generate a
594 (Also dividing the most negative integer by \-1 may generate
596 Ignoring this signal might lead to an endless loop.
598 POSIX.1-1990 disallowed setting the action for
602 POSIX.1-2001 allows this possibility, so that ignoring
604 can be used to prevent the creation of zombies (see
606 Nevertheless, the historical BSD and System V behaviors for ignoring
608 differ, so that the only completely portable method of ensuring that
609 terminated children do not become zombies is to catch the
615 POSIX.1-1990 only specified
623 Use of these latter values in
625 may be less portable in applications intended for older
626 Unix implementations.
630 flag is compatible with the SVr4 flag of the same name.
634 flag is compatible with the SVr4 flag of the same name under kernels
636 On older kernels the Linux implementation
637 allowed the receipt of any signal, not just the one we are installing
638 (effectively overriding any
643 can be called with a null second argument to query the current signal
645 It can also be used to check whether a given signal is valid for
646 the current machine by calling it with null second and third arguments.
648 It is not possible to block
649 .BR SIGKILL " or " SIGSTOP
650 (by specifying them in
652 Attempts to do so are silently ignored.
656 for details on manipulating signal sets.
660 for a list of the async-signal-safe functions that can be
661 safely called inside from inside a signal handler.
663 Before the introduction of
665 it was also possible to get some additional information,
668 with second argument of type
669 .IR "struct sigcontext".
670 See the relevant kernel sources for details.
671 This use is obsolete now.
673 In kernels up to and including 2.6.13, specifying
677 prevents not only the delivered signal from being masked during
678 execution of the handler, but also the signals specified in
680 This bug was fixed in kernel 2.6.14.
698 .BR siginterrupt (3),