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16 .\" manual page may be incorrect or out-of-date.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.
42 .\" 2010-06-11 Andi Kleen, add hwpoison signal extensions
43 .\" 2010-06-11 mtk, improvements to discussion of various siginfo_t fields.
45 .TH SIGACTION 2 2012-04-26 "Linux" "Linux Programmer's Manual"
47 sigaction \- examine and change a signal action
50 .B #include <signal.h>
52 .BI "int sigaction(int " signum ", const struct sigaction *" act ,
53 .BI " struct sigaction *" oldact );
57 Feature Test Macro Requirements for glibc (see
58 .BR feature_test_macros (7)):
63 _POSIX_C_SOURCE\ >=\ 1 || _XOPEN_SOURCE || _POSIX_SOURCE
68 system call is used to change the action taken by a process on
69 receipt of a specific signal.
72 for an overview of signals.)
75 specifies the signal and can be any valid signal except
82 is non-NULL, the new action for signal
88 is non-NULL, the previous action is saved in
93 structure is defined as something like:
98 void (*sa_handler)(int);
99 void (*sa_sigaction)(int, siginfo_t *, void *);
102 void (*sa_restorer)(void);
107 On some architectures a union is involved: do not assign to both
114 element is obsolete and should not be used.
115 POSIX does not specify a
120 specifies the action to be associated with
124 for the default action,
126 to ignore this signal, or a pointer to a signal handling function.
127 This function receives the signal number as its only argument.
137 specifies the signal-handling function for
139 This function receives the signal number as its first argument, a
142 as its second argument and a pointer to a
144 (cast to \fIvoid\ *\fP) as its third argument.
145 (Commonly, the handler function doesn't make any use of the third argument.
148 for further information about
152 specifies a mask of signals which should be blocked
153 (i.e., added to the signal mask of the thread in which
154 the signal handler is invoked)
155 during execution of the signal handler.
156 In addition, the signal which triggered the handler
157 will be blocked, unless the
162 specifies a set of flags which modify the behavior of the signal.
163 It is formed by the bitwise OR of zero or more of the following:
171 do not receive notification when child processes stop (i.e., when they
173 .BR SIGSTOP ", " SIGTSTP ", " SIGTTIN
176 or resume (i.e., they receive
180 This flag is only meaningful when establishing a handler for
183 .BR SA_NOCLDWAIT " (since Linux 2.6)"
184 .\" To be precise: Linux 2.5.60 -- MTK
189 do not transform children into zombies when they terminate.
192 This flag is only meaningful when establishing a handler for
194 or when setting that signal's disposition to
199 flag is set when establishing a handler for
201 POSIX.1 leaves it unspecified whether a
203 signal is generated when a child process terminates.
206 signal is generated in this case;
207 on some other implementations, it is not.
210 Do not prevent the signal from being received from within its own signal
212 This flag is only meaningful when establishing a signal handler.
214 is an obsolete, nonstandard synonym for this flag.
217 Call the signal handler on an alternate signal stack provided by
219 If an alternate stack is not available, the default stack will be used.
220 This flag is only meaningful when establishing a signal handler.
223 Restore the signal action to the default state once the signal handler
225 This flag is only meaningful when establishing a signal handler.
227 is an obsolete, nonstandard synonym for this flag.
230 Provide behavior compatible with BSD signal semantics by making certain
231 system calls restartable across signals.
232 This flag is only meaningful when establishing a signal handler.
235 for a discussion of system call restarting.
237 .BR SA_SIGINFO " (since Linux 2.2)"
238 The signal handler takes 3 arguments, not one.
241 should be set instead of
243 This flag is only meaningful when establishing a signal handler.
246 .\" field was added in Linux 2.1.86.)
253 is a struct with the following elements:
258 int si_signo; /* Signal number */
259 int si_errno; /* An errno value */
260 int si_code; /* Signal code */
261 int si_trapno; /* Trap number that caused
262 hardware-generated signal
263 (unused on most architectures) */
265 .\" si_trapno seems to be only used on SPARC and Alpha;
266 .\" this page could use a little more detail on its purpose there.
267 pid_t si_pid; /* Sending process ID */
268 uid_t si_uid; /* Real user ID of sending process */
269 int si_status; /* Exit value or signal */
270 clock_t si_utime; /* User time consumed */
271 clock_t si_stime; /* System time consumed */
272 sigval_t si_value; /* Signal value */
273 int si_int; /* POSIX.1b signal */
274 void *si_ptr; /* POSIX.1b signal */
275 int si_overrun; /* Timer overrun count; POSIX.1b timers */
276 int si_timerid; /* Timer ID; POSIX.1b timers */
277 .\" In the kernel: si_tid
278 void *si_addr; /* Memory location which caused fault */
279 long si_band; /* Band event (was \fIint\fP in
280 glibc 2.3.2 and earlier) */
281 int si_fd; /* File descriptor */
282 short si_addr_lsb; /* Least significant bit of address
283 (since kernel 2.6.32) */
288 .IR si_signo ", " si_errno " and " si_code
289 are defined for all signals.
291 is generally unused on Linux.)
292 The rest of the struct may be a union, so that one should only
293 read the fields that are meaningful for the given signal:
300 .IR si_pid " and " si_uid .
301 In addition, signals sent with
304 .IR si_int " and " si_ptr
305 with the values specified by the sender of the signal;
310 Signals sent by POSIX.1b timers (since Linux 2.6) fill in
316 field is an internal ID used by the kernel to identify
317 the timer; it is not the same as the timer ID returned by
318 .BR timer_create (2).
321 field is the timer overrun count;
322 this is the same information as is obtained by a call to
323 .BR timer_getoverrun (2).
324 These fields are nonstandard Linux extensions.
326 Signals sent for message queue notification (see the description of
331 .IR si_int / si_ptr ,
337 with the process ID of the message sender; and
339 with the real user ID of the message sender.
343 .IR si_pid ", " si_uid ", " si_status ", " si_utime " and " si_stime ,
344 providing information about the child.
347 field is the process ID of the child;
349 is the child's real user ID.
352 field contains the exit status of the child (if
356 or the signal number that caused the process to change state.
361 contain the user and system CPU time used by the child process;
362 these fields do not include the times used by waited-for children (unlike
366 In kernels up to 2.6, and since 2.6.27, these fields report
368 .IR sysconf(_SC_CLK_TCK) .
369 In 2.6 kernels before 2.6.27,
370 a bug meant that these fields reported time in units
371 of the (configurable) system jiffy (see
374 .\" When si_utime and si_stime where originally implemented, the
375 .\" measurement unit was HZ, which was the same as clock ticks
376 .\" (sysconf(_SC_CLK_TCK)). In 2.6, HZ became configurable, and
377 .\" was *still* used as the unit to return the info these fields,
378 .\" with the result that the field values depended on the the
379 .\" configured HZ. Of course, the should have been measured in
380 .\" USER_HZ instead, so that sysconf(_SC_CLK_TCK) could be used to
381 .\" convert to seconds. I have a queued patch to fix this:
382 .\" http://thread.gmane.org/gmane.linux.kernel/698061/ .
383 .\" This patch made it into 2.6.27.
384 .\" But note that these fields still don't return the times of
385 .\" waited-for children (as is done by getrusage() and times()
386 .\" and wait4()). Solaris 8 does include child times.
396 with the address of the fault.
397 .\" FIXME SIGTRAP also sets the following for ptrace_notify() ?
398 .\" info.si_code = exit_code;
399 .\" info.si_pid = task_pid_vnr(current);
400 .\" info.si_uid = current_uid(); /* Real UID */
401 On some architectures,
402 these signals also fill in the
413 This field indicates the least significant bit of the reported address
414 and therefore the extent of the corruption.
415 For example, if a full page was corrupted,
418 .IR log2(sysconf(_SC_PAGESIZE)) .
422 are Linux-specific extensions.
425 (the two names are synonyms on Linux)
427 .IR si_band " and " si_fd .
430 event is a bit mask containing the same values as are filled in the
436 field indicates the file descriptor for which the I/O event occurred.
439 is a value (not a bit mask)
440 indicating why this signal was sent.
441 The following list shows the values which can be placed in
443 for any signal, along with reason that the signal was generated.
459 POSIX message queue state changed (since Linux 2.6.6); see
468 (only in kernels up to Linux 2.2; from Linux 2.4 onward
479 .\" SI_DETHREAD is defined in 2.6.9 sources, but isn't implemented
480 .\" It appears to have been an idea that was tried during 2.5.6
481 .\" through to 2.5.24 and then was backed out.
484 The following values can be placed in
498 illegal addressing mode
516 The following values can be placed in
524 integer divide by zero
530 floating-point divide by zero
533 floating-point overflow
536 floating-point underflow
539 floating-point inexact result
542 floating-point invalid operation
545 subscript out of range
548 The following values can be placed in
556 address not mapped to object
559 invalid permissions for mapped object
562 The following values can be placed in
570 invalid address alignment
573 nonexistent physical address
576 object-specific hardware error
578 .BR BUS_MCEERR_AR " (since Linux 2.6.32)"
579 Hardware memory error consumed on a machine check; action required.
581 .BR BUS_MCEERR_AO " (since Linux 2.6.32)"
582 Hardware memory error detected in process but not consumed; action optional.
585 The following values can be placed in
598 .BR TRAP_BRANCH " (since Linux 2.4)"
599 process taken branch trap
601 .BR TRAP_HWBKPT " (since Linux 2.4)"
602 hardware breakpoint/watchpoint
605 The following values can be placed in
619 child terminated abnormally
622 traced child has trapped
628 stopped child has continued (since Linux 2.6.9)
631 The following values can be placed in
642 output buffers available
645 input message available
651 high priority input available
658 returns 0 on success and \-1 on error.
662 .IR act " or " oldact
663 points to memory which is not a valid part of the process address space.
666 An invalid signal was specified.
667 This will also be generated if an attempt
668 is made to change the action for
669 .BR SIGKILL " or " SIGSTOP ", "
670 which cannot be caught or ignored.
673 .\" SVr4 does not document the EINTR condition.
677 inherits a copy of its parent's signal dispositions.
680 the dispositions of handled signals are reset to the default;
681 the dispositions of ignored signals are left unchanged.
683 According to POSIX, the behavior of a process is undefined after it
689 signal that was not generated by
693 Integer division by zero has undefined result.
694 On some architectures it will generate a
697 (Also dividing the most negative integer by \-1 may generate
699 Ignoring this signal might lead to an endless loop.
701 POSIX.1-1990 disallowed setting the action for
705 POSIX.1-2001 allows this possibility, so that ignoring
707 can be used to prevent the creation of zombies (see
709 Nevertheless, the historical BSD and System V behaviors for ignoring
711 differ, so that the only completely portable method of ensuring that
712 terminated children do not become zombies is to catch the
718 POSIX.1-1990 only specified
726 Use of these latter values in
728 may be less portable in applications intended for older
729 UNIX implementations.
733 flag is compatible with the SVr4 flag of the same name.
737 flag is compatible with the SVr4 flag of the same name under kernels
739 On older kernels the Linux implementation
740 allowed the receipt of any signal, not just the one we are installing
741 (effectively overriding any
746 can be called with a NULL second argument to query the current signal
748 It can also be used to check whether a given signal is valid for
749 the current machine by calling it with NULL second and third arguments.
751 It is not possible to block
752 .BR SIGKILL " or " SIGSTOP
753 (by specifying them in
755 Attempts to do so are silently ignored.
759 for details on manipulating signal sets.
763 for a list of the async-signal-safe functions that can be
764 safely called inside from inside a signal handler.
766 Before the introduction of
768 it was also possible to get some additional information,
771 with second argument of type
772 .IR "struct sigcontext".
773 See the relevant kernel sources for details.
774 This use is obsolete now.
776 In kernels up to and including 2.6.13, specifying
780 prevents not only the delivered signal from being masked during
781 execution of the handler, but also the signals specified in
783 This bug was fixed in kernel 2.6.14.
800 .BR siginterrupt (3),