1KILL(P) POSIX Programmer's Manual KILL(P)
2
6 kill - send a signal to a process or a group of processes
7
9 #include <signal.h>
10 int kill(pid_t pid, int sig);
12
15 The kill() function shall send a signal to a process or a group of pro‐
16 cesses specified by pid. The signal to be sent is specified by sig and
17 is either one from the list given in <signal.h> or 0. If sig is 0 (the
18 null signal), error checking is performed but no signal is actually
19 sent. The null signal can be used to check the validity of pid.
20 For a process to have permission to send a signal to a process desig‐
22 nated by pid, unless the sending process has appropriate privileges,
23 the real or effective user ID of the sending process shall match the
24 real or saved set-user-ID of the receiving process.
25 If pid is greater than 0, sig shall be sent to the process whose
27 process ID is equal to pid.
28 If pid is 0, sig shall be sent to all processes (excluding an unspeci‐
30 fied set of system processes) whose process group ID is equal to the
31 process group ID of the sender, and for which the process has permis‐
32 sion to send a signal.
33 If pid is -1, sig shall be sent to all processes (excluding an unspeci‐
35 fied set of system processes) for which the process has permission to
36 send that signal.
37 If pid is negative, but not -1, sig shall be sent to all processes
39 (excluding an unspecified set of system processes) whose process group
40 ID is equal to the absolute value of pid, and for which the process has
41 permission to send a signal.
42 If the value of pid causes sig to be generated for the sending process,
44 and if sig is not blocked for the calling thread and if no other thread
45 has sig unblocked or is waiting in a sigwait() function for sig, either
46 sig or at least one pending unblocked signal shall be delivered to the
47 sending thread before kill() returns.
48 The user ID tests described above shall not be applied when sending
50 SIGCONT to a process that is a member of the same session as the send‐
51 ing process.
52 An implementation that provides extended security controls may impose
54 further implementation-defined restrictions on the sending of signals,
55 including the null signal. In particular, the system may deny the exis‐
56 tence of some or all of the processes specified by pid.
57 The kill() function is successful if the process has permission to send
59 sig to any of the processes specified by pid. If kill() fails, no sig‐
60 nal shall be sent.
61
63 Upon successful completion, 0 shall be returned. Otherwise, -1 shall be
64 returned and errno set to indicate the error.
65
67 The kill() function shall fail if:
68 EINVAL The value of the sig argument is an invalid or unsupported sig‐
70 nal number.
71 EPERM The process does not have permission to send the signal to any
73 receiving process.
74 ESRCH No process or process group can be found corresponding to that
76 specified by pid.
77 The following sections are informative.
80
82 None.
83
85 None.
86
88 The semantics for permission checking for kill() differed between Sys‐
89 tem V and most other implementations, such as Version 7 or 4.3 BSD. The
90 semantics chosen for this volume of IEEE Std 1003.1-2001 agree with
91 System V. Specifically, a set-user-ID process cannot protect itself
92 against signals (or at least not against SIGKILL) unless it changes its
93 real user ID. This choice allows the user who starts an application to
94 send it signals even if it changes its effective user ID. The other
95 semantics give more power to an application that wants to protect
96 itself from the user who ran it.
97 Some implementations provide semantic extensions to the kill() function
99 when the absolute value of pid is greater than some maximum, or other‐
100 wise special, value. Negative values are a flag to kill(). Since most
101 implementations return [ESRCH] in this case, this behavior is not
102 included in this volume of IEEE Std 1003.1-2001, although a conforming
103 implementation could provide such an extension.
104 The implementation-defined processes to which a signal cannot be sent
106 may include the scheduler or init.
107 There was initially strong sentiment to specify that, if pid specifies
109 that a signal be sent to the calling process and that signal is not
110 blocked, that signal would be delivered before kill() returns. This
111 would permit a process to call kill() and be guaranteed that the call
112 never return. However, historical implementations that provide only the
113 signal() function make only the weaker guarantee in this volume of
114 IEEE Std 1003.1-2001, because they only deliver one signal each time a
115 process enters the kernel. Modifications to such implementations to
116 support the sigaction() function generally require entry to the kernel
117 following return from a signal-catching function, in order to restore
118 the signal mask. Such modifications have the effect of satisfying the
119 stronger requirement, at least when sigaction() is used, but not neces‐
120 sarily when signal() is used. The developers of this volume of
121 IEEE Std 1003.1-2001 considered making the stronger requirement except
122 when signal() is used, but felt this would be unnecessarily complex.
123 Implementors are encouraged to meet the stronger requirement whenever
124 possible. In practice, the weaker requirement is the same, except in
125 the rare case when two signals arrive during a very short window. This
126 reasoning also applies to a similar requirement for sigprocmask().
127 In 4.2 BSD, the SIGCONT signal can be sent to any descendant process
129 regardless of user-ID security checks. This allows a job control shell
130 to continue a job even if processes in the job have altered their user
131 IDs (as in the su command). In keeping with the addition of the concept
132 of sessions, similar functionality is provided by allowing the SIGCONT
133 signal to be sent to any process in the same session regardless of user
134 ID security checks. This is less restrictive than BSD in the sense
135 that ancestor processes (in the same session) can now be the recipient.
136 It is more restrictive than BSD in the sense that descendant processes
137 that form new sessions are now subject to the user ID checks. A similar
138 relaxation of security is not necessary for the other job control sig‐
139 nals since those signals are typically sent by the terminal driver in
140 recognition of special characters being typed; the terminal driver
141 bypasses all security checks.
142 In secure implementations, a process may be restricted from sending a
144 signal to a process having a different security label. In order to pre‐
145 vent the existence or nonexistence of a process from being used as a
146 covert channel, such processes should appear nonexistent to the sender;
147 that is, [ESRCH] should be returned, rather than [EPERM], if pid refers
148 only to such processes.
149 Existing implementations vary on the result of a kill() with pid indi‐
151 cating an inactive process (a terminated process that has not been
152 waited for by its parent). Some indicate success on such a call (sub‐
153 ject to permission checking), while others give an error of [ESRCH].
154 Since the definition of process lifetime in this volume of
155 IEEE Std 1003.1-2001 covers inactive processes, the [ESRCH] error as
156 described is inappropriate in this case. In particular, this means that
157 an application cannot have a parent process check for termination of a
158 particular child with kill(). (Usually this is done with the null sig‐
159 nal; this can be done reliably with waitpid().)
160 There is some belief that the name kill() is misleading, since the
162 function is not always intended to cause process termination. However,
163 the name is common to all historical implementations, and any change
164 would be in conflict with the goal of minimal changes to existing
165 application code.
166
168 None.
169
171 getpid() , raise() , setsid() , sigaction() , sigqueue() , the Base
172 Definitions volume of IEEE Std 1003.1-2001, <signal.h>, <sys/types.h>
173
175 Portions of this text are reprinted and reproduced in electronic form
176 from IEEE Std 1003.1, 2003 Edition, Standard for Information Technology
177 -- Portable Operating System Interface (POSIX), The Open Group Base
178 Specifications Issue 6, Copyright (C) 2001-2003 by the Institute of
179 Electrical and Electronics Engineers, Inc and The Open Group. In the
180 event of any discrepancy between this version and the original IEEE and
181 The Open Group Standard, the original IEEE and The Open Group Standard
182 is the referee document. The original Standard can be obtained online
183 at http://www.opengroup.org/unix/online.html .
184IEEE/The Open Group 2003 KILL(P)