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16 .\" manual page may be incorrect or out-of-date.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 2008-07-08 "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 );
56 system call is used to change the action taken by a process on
57 receipt of a specific signal.
60 specifies the signal and can be any valid signal except
67 is non-null, the new action for signal
73 is non-null, the previous action is saved in
78 structure is defined as something like:
83 void (*sa_handler)(int);
84 void (*sa_sigaction)(int, siginfo_t *, void *);
87 void (*sa_restorer)(void);
92 On some architectures a union is involved: do not assign to both
99 element is obsolete and should not be used.
100 POSIX does not specify a
105 specifies the action to be associated with
109 for the default action,
111 to ignore this signal, or a pointer to a signal handling function.
112 This function receives the signal number as its only argument.
122 specifies the signal-handling function for
124 This function receives the signal number as its first argument, a
127 as its second argument and a pointer to a
129 (cast to \fIvoid\ *\fP) as its third argument.
132 gives a mask of signals which should be blocked during execution of
134 In addition, the signal which triggered the handler
135 will be blocked, unless the
140 specifies a set of flags which modify the behavior of the signal.
141 It is formed by the bitwise OR of zero or more of the following:
149 do not receive notification when child processes stop (i.e., when they
151 .BR SIGSTOP ", " SIGTSTP ", " SIGTTIN
154 or resume (i.e., they receive
158 This flag is only meaningful when establishing a handler for
161 .BR SA_NOCLDWAIT " (Since Linux 2.6)"
162 .\" To be precise: Linux 2.5.60 -- MTK
167 do not transform children into zombies when they terminate.
170 This flag is only meaningful when establishing a handler for
172 or when setting that signal's disposition to
177 flag is set when establishing a handler for
179 POSIX.1 leaves it unspecified whether a
181 signal is generated when a child process terminates.
184 signal is generated in this case;
185 on some other implementations, it is not.
188 Do not prevent the signal from being received from within its own signal
190 This flag is only meaningful when establishing a signal handler.
192 is an obsolete, non-standard synonym for this flag.
195 Call the signal handler on an alternate signal stack provided by
197 If an alternate stack is not available, the default stack will be used.
198 This flag is only meaningful when establishing a signal handler.
201 Restore the signal action to the default state once the signal handler
203 This flag is only meaningful when establishing a signal handler.
205 is an obsolete, non-standard synonym for this flag.
208 Provide behavior compatible with BSD signal semantics by making certain
209 system calls restartable across signals.
210 This flag is only meaningful when establishing a signal handler.
213 for a discussion of system call restarting.
215 .BR SA_SIGINFO " (since Linux 2.2)"
216 The signal handler takes 3 arguments, not one.
219 should be set instead of
221 This flag is only meaningful when establishing a signal handler.
224 .\" field was added in Linux 2.1.86.)
231 is a struct with the following elements:
236 int si_signo; /* Signal number */
237 int si_errno; /* An errno value */
238 int si_code; /* Signal code */
239 int si_trapno; /* Trap number that caused
240 hardware-generated signal
241 (unused on most architectures) */
243 .\" si_trapno seems to be only used on SPARC and Alpha;
244 .\" this page could use a little more detail on its purpose there.
245 pid_t si_pid; /* Sending process ID */
246 uid_t si_uid; /* Real user ID of sending process */
247 int si_status; /* Exit value or signal */
248 clock_t si_utime; /* User time consumed */
249 clock_t si_stime; /* System time consumed */
250 sigval_t si_value; /* Signal value */
251 int si_int; /* POSIX.1b signal */
252 void *si_ptr; /* POSIX.1b signal */
253 int si_overrun; /* Timer overrun count; POSIX.1b timers */
254 int si_timerid; /* Timer ID; POSIX.1b timers */
255 .\" In the kernel: si_tid
256 void *si_addr; /* Memory location which caused fault */
257 int si_band; /* Band event */
258 int si_fd; /* File descriptor */
263 .IR si_signo ", " si_errno " and " si_code
264 are defined for all signals.
266 is generally unused on Linux.)
267 The rest of the struct may be a union, so that one should only
268 read the fields that are meaningful for the given signal:
273 .IR si_pid " and " si_uid .
275 POSIX.1b timers (since Linux 2.6) fill in
281 field is an internal ID used by the kernel to identify
282 the timer; it is not the same as the timer ID returned by
283 .BR timer_create (3).
287 .IR si_status ", " si_utime " and " si_stime .
289 .\" When si_utime and si_stime where originally implemented, the
290 .\" measurement unit was HZ, which was the same as clock ticks
291 .\" (sysconf(_SC_CLK_TCK)). In 2.6, HZ became configurable, and
292 .\" was *still* used as the unit to return the info these fields,
293 .\" with the result that the field values depended on the the
294 .\" configured HZ. Of course, the should have been measured in
295 .\" USER_HZ instead, so that sysconf(_SC_CLK_TCK) could be used to
296 .\" convert to seconds. I have a queued patch to fix this:
297 .\" http://thread.gmane.org/gmane.linux.kernel/698061/ .
298 .\" Maybe it will make it into 2.6.27.
299 .\" But note that these fields still don't return the times of
300 .\" waited for children (as is done by getrusage() and times()
301 .\" and wait4()). Solaris 8 does include child times.
303 .IR si_int " and " si_ptr
304 are specified by the sender of the POSIX.1b signal.
316 with the address of the fault.
319 .IR si_band " and " si_fd .
322 is a value (not a bit mask)
323 indicating why this signal was sent.
324 The following list shows the values which can be placed in
326 for any signal, along with reason that the signal was generated.
344 POSIX message queue state changed (since Linux 2.6.6); see
358 .\" SI_DETHREAD is defined in 2.6.9 sources, but isn't implemented
359 .\" It appears to have been an idea that was tried during 2.5.6
360 .\" through to 2.5.24 and then was backed out.
363 The following values can be placed in
377 illegal addressing mode
395 The following values can be placed in
403 integer divide by zero
409 floating point divide by zero
412 floating point overflow
415 floating point underflow
418 floating point inexact result
421 floating point invalid operation
424 subscript out of range
427 The following values can be placed in
435 address not mapped to object
438 invalid permissions for mapped object
441 The following values can be placed in
449 invalid address alignment
452 nonexistent physical address
455 object-specific hardware error
458 The following values can be placed in
472 The following values can be placed in
486 child terminated abnormally
489 traced child has trapped
495 stopped child has continued (since Linux 2.6.9)
498 The following values can be placed in
509 output buffers available
512 input message available
518 high priority input available
525 returns 0 on success and \-1 on error.
529 .IR act " or " oldact
530 points to memory which is not a valid part of the process address space.
533 An invalid signal was specified.
534 This will also be generated if an attempt
535 is made to change the action for
536 .BR SIGKILL " or " SIGSTOP ", "
537 which cannot be caught or ignored.
540 .\" SVr4 does not document the EINTR condition.
543 According to POSIX, the behavior of a process is undefined after it
549 signal that was not generated by
553 Integer division by zero has undefined result.
554 On some architectures it will generate a
557 (Also dividing the most negative integer by \-1 may generate
559 Ignoring this signal might lead to an endless loop.
561 POSIX.1-1990 disallowed setting the action for
565 POSIX.1-2001 allows this possibility, so that ignoring
567 can be used to prevent the creation of zombies (see
569 Nevertheless, the historical BSD and System V behaviors for ignoring
571 differ, so that the only completely portable method of ensuring that
572 terminated children do not become zombies is to catch the
578 POSIX.1-1990 only specified
586 Use of these latter values in
588 may be less portable in applications intended for older
589 Unix implementations.
593 flag is compatible with the SVr4 flag of the same name.
597 flag is compatible with the SVr4 flag of the same name under kernels
599 On older kernels the Linux implementation
600 allowed the receipt of any signal, not just the one we are installing
601 (effectively overriding any
606 can be called with a null second argument to query the current signal
608 It can also be used to check whether a given signal is valid for
609 the current machine by calling it with null second and third arguments.
611 It is not possible to block
612 .BR SIGKILL " or " SIGSTOP
613 (by specifying them in
615 Attempts to do so are silently ignored.
619 for details on manipulating signal sets.
623 for a list of the async-signal-safe functions that can be
624 safely called inside from inside a signal handler.
626 Before the introduction of
628 it was also possible to get some additional information,
631 with second argument of type
632 .IR "struct sigcontext".
633 See the relevant kernel sources for details.
634 This use is obsolete now.
636 In kernels up to and including 2.6.13, specifying
640 prevents not only the delivered signal from being masked during
641 execution of the handler, but also the signals specified in
643 This bug was fixed in kernel 2.6.14.
661 .BR siginterrupt (3),