1SEMCTL(2) Linux Programmer's Manual SEMCTL(2)
2
3
4
6 semctl - System V semaphore control operations
7
9 #include <sys/types.h>
10 #include <sys/ipc.h>
11 #include <sys/sem.h>
12
13 int semctl(int semid, int semnum, int cmd, ...);
14
16 semctl() performs the control operation specified by cmd on the Sys‐
17 tem V semaphore set identified by semid, or on the semnum-th semaphore
18 of that set. (The semaphores in a set are numbered starting at 0.)
19
20 This function has three or four arguments, depending on cmd. When
21 there are four, the fourth has the type union semun. The calling pro‐
22 gram must define this union as follows:
23
24 union semun {
25 int val; /* Value for SETVAL */
26 struct semid_ds *buf; /* Buffer for IPC_STAT, IPC_SET */
27 unsigned short *array; /* Array for GETALL, SETALL */
28 struct seminfo *__buf; /* Buffer for IPC_INFO
29 (Linux-specific) */
30 };
31
32 The semid_ds data structure is defined in <sys/sem.h> as follows:
33
34 struct semid_ds {
35 struct ipc_perm sem_perm; /* Ownership and permissions */
36 time_t sem_otime; /* Last semop time */
37 time_t sem_ctime; /* Creation time/time of last
38 modification via semctl() */
39 unsigned long sem_nsems; /* No. of semaphores in set */
40 };
41
42 The fields of the semid_ds structure are as follows:
43
44 sem_perm This is an ipc_perm structure (see below) that specifies the
45 access permissions on the semaphore set.
46
47 sem_otime Time of last semop(2) system call.
48
49 sem_ctime Time of creation of semaphore set or time of last semctl()
50 IPCSET, SETVAL, or SETALL operation.
51
52 sem_nsems Number of semaphores in the set. Each semaphore of the set
53 is referenced by a nonnegative integer ranging from 0 to
54 sem_nsems-1.
55
56 The ipc_perm structure is defined as follows (the highlighted fields
57 are settable using IPC_SET):
58
59 struct ipc_perm {
60 key_t __key; /* Key supplied to semget(2) */
61 uid_t uid; /* Effective UID of owner */
62 gid_t gid; /* Effective GID of owner */
63 uid_t cuid; /* Effective UID of creator */
64 gid_t cgid; /* Effective GID of creator */
65 unsigned short mode; /* Permissions */
66 unsigned short __seq; /* Sequence number */
67 };
68
69 The least significant 9 bits of the mode field of the ipc_perm struc‐
70 ture define the access permissions for the shared memory segment. The
71 permission bits are as follows:
72
73 0400 Read by user
74 0200 Write by user
75 0040 Read by group
76 0020 Write by group
77 0004 Read by others
78 0002 Write by others
79
80 In effect, "write" means "alter" for a semaphore set. Bits 0100, 0010,
81 and 0001 (the execute bits) are unused by the system.
82
83 Valid values for cmd are:
84
85 IPC_STAT
86 Copy information from the kernel data structure associated with
87 semid into the semid_ds structure pointed to by arg.buf. The
88 argument semnum is ignored. The calling process must have read
89 permission on the semaphore set.
90
91 IPC_SET
92 Write the values of some members of the semid_ds structure
93 pointed to by arg.buf to the kernel data structure associated
94 with this semaphore set, updating also its sem_ctime member.
95
96 The following members of the structure are updated:
97 sem_perm.uid, sem_perm.gid, and (the least significant 9 bits
98 of) sem_perm.mode.
99
100 The effective UID of the calling process must match the owner
101 (sem_perm.uid) or creator (sem_perm.cuid) of the semaphore set,
102 or the caller must be privileged. The argument semnum is ig‐
103 nored.
104
105 IPC_RMID
106 Immediately remove the semaphore set, awakening all processes
107 blocked in semop(2) calls on the set (with an error return and
108 errno set to EIDRM). The effective user ID of the calling
109 process must match the creator or owner of the semaphore set, or
110 the caller must be privileged. The argument semnum is ignored.
111
112 IPC_INFO (Linux-specific)
113 Return information about system-wide semaphore limits and param‐
114 eters in the structure pointed to by arg.__buf. This structure
115 is of type seminfo, defined in <sys/sem.h> if the _GNU_SOURCE
116 feature test macro is defined:
117
118 struct seminfo {
119 int semmap; /* Number of entries in semaphore
120 map; unused within kernel */
121 int semmni; /* Maximum number of semaphore sets */
122 int semmns; /* Maximum number of semaphores in all
123 semaphore sets */
124 int semmnu; /* System-wide maximum number of undo
125 structures; unused within kernel */
126 int semmsl; /* Maximum number of semaphores in a
127 set */
128 int semopm; /* Maximum number of operations for
129 semop(2) */
130 int semume; /* Maximum number of undo entries per
131 process; unused within kernel */
132 int semusz; /* Size of struct sem_undo */
133 int semvmx; /* Maximum semaphore value */
134 int semaem; /* Max. value that can be recorded for
135 semaphore adjustment (SEM_UNDO) */
136 };
137
138 The semmsl, semmns, semopm, and semmni settings can be changed
139 via /proc/sys/kernel/sem; see proc(5) for details.
140
141 SEM_INFO (Linux-specific)
142 Return a seminfo structure containing the same information as
143 for IPC_INFO, except that the following fields are returned with
144 information about system resources consumed by semaphores: the
145 semusz field returns the number of semaphore sets that currently
146 exist on the system; and the semaem field returns the total num‐
147 ber of semaphores in all semaphore sets on the system.
148
149 SEM_STAT (Linux-specific)
150 Return a semid_ds structure as for IPC_STAT. However, the semid
151 argument is not a semaphore identifier, but instead an index
152 into the kernel's internal array that maintains information
153 about all semaphore sets on the system.
154
155 SEM_STAT_ANY (Linux-specific, since Linux 4.17)
156 Return a seminfo structure containing the same information as
157 for SEM_STAT. However, sem_perm.mode is not checked for read
158 access for semid meaning that any user can employ this operation
159 (just as any user may read /proc/sysvipc/sem to obtain the same
160 information).
161
162 GETALL Return semval (i.e., the current value) for all semaphores of
163 the set into arg.array. The argument semnum is ignored. The
164 calling process must have read permission on the semaphore set.
165
166 GETNCNT
167 Return the semncnt value for the semnum-th semaphore of the set
168 (i.e., the number of processes waiting for the semaphore's value
169 to increase). The calling process must have read permission on
170 the semaphore set.
171
172 GETPID Return the sempid value for the semnum-th semaphore of the set.
173 This is the PID of the process that last performed an operation
174 on that semaphore (but see NOTES). The calling process must
175 have read permission on the semaphore set.
176
177 GETVAL Return semval (i.e., the semaphore value) for the semnum-th sem‐
178 aphore of the set. The calling process must have read permis‐
179 sion on the semaphore set.
180
181 GETZCNT
182 Return the semzcnt value for the semnum-th semaphore of the set
183 (i.e., the number of processes waiting for the semaphore value
184 to become 0). The calling process must have read permission on
185 the semaphore set.
186
187 SETALL Set the semval values for all semaphores of the set using
188 arg.array, updating also the sem_ctime member of the semid_ds
189 structure associated with the set. Undo entries (see semop(2))
190 are cleared for altered semaphores in all processes. If the
191 changes to semaphore values would permit blocked semop(2) calls
192 in other processes to proceed, then those processes are woken
193 up. The argument semnum is ignored. The calling process must
194 have alter (write) permission on the semaphore set.
195
196 SETVAL Set the semaphore value (semval) to arg.val for the semnum-th
197 semaphore of the set, updating also the sem_ctime member of the
198 semid_ds structure associated with the set. Undo entries are
199 cleared for altered semaphores in all processes. If the changes
200 to semaphore values would permit blocked semop(2) calls in other
201 processes to proceed, then those processes are woken up. The
202 calling process must have alter permission on the semaphore set.
203
205 On failure, semctl() returns -1 with errno indicating the error.
206
207 Otherwise, the system call returns a nonnegative value depending on cmd
208 as follows:
209
210 GETNCNT
211 the value of semncnt.
212
213 GETPID the value of sempid.
214
215 GETVAL the value of semval.
216
217 GETZCNT
218 the value of semzcnt.
219
220 IPC_INFO
221 the index of the highest used entry in the kernel's internal ar‐
222 ray recording information about all semaphore sets. (This in‐
223 formation can be used with repeated SEM_STAT or SEM_STAT_ANY op‐
224 erations to obtain information about all semaphore sets on the
225 system.)
226
227 SEM_INFO
228 as for IPC_INFO.
229
230 SEM_STAT
231 the identifier of the semaphore set whose index was given in
232 semid.
233
234 SEM_STAT_ANY
235 as for SEM_STAT.
236
237 All other cmd values return 0 on success.
238
240 On failure, errno will be set to one of the following:
241
242 EACCES The argument cmd has one of the values GETALL, GETPID, GETVAL,
243 GETNCNT, GETZCNT, IPC_STAT, SEM_STAT, SEM_STAT_ANY, SETALL, or
244 SETVAL and the calling process does not have the required per‐
245 missions on the semaphore set and does not have the
246 CAP_IPC_OWNER capability in the user namespace that governs its
247 IPC namespace.
248
249 EFAULT The address pointed to by arg.buf or arg.array isn't accessible.
250
251 EIDRM The semaphore set was removed.
252
253 EINVAL Invalid value for cmd or semid. Or: for a SEM_STAT operation,
254 the index value specified in semid referred to an array slot
255 that is currently unused.
256
257 EPERM The argument cmd has the value IPC_SET or IPC_RMID but the ef‐
258 fective user ID of the calling process is not the creator (as
259 found in sem_perm.cuid) or the owner (as found in sem_perm.uid)
260 of the semaphore set, and the process does not have the
261 CAP_SYS_ADMIN capability.
262
263 ERANGE The argument cmd has the value SETALL or SETVAL and the value to
264 which semval is to be set (for some semaphore of the set) is
265 less than 0 or greater than the implementation limit SEMVMX.
266
268 POSIX.1-2001, POSIX.1-2008, SVr4.
269
270 POSIX.1 specifies the sem_nsems field of the semid_ds structure as hav‐
271 ing the type unsigned short, and the field is so defined on most other
272 systems. It was also so defined on Linux 2.2 and earlier, but, since
273 Linux 2.4, the field has the type unsigned long.
274
276 The inclusion of <sys/types.h> and <sys/ipc.h> isn't required on Linux
277 or by any version of POSIX. However, some old implementations required
278 the inclusion of these header files, and the SVID also documented their
279 inclusion. Applications intended to be portable to such old systems
280 may need to include these header files.
281
282 The IPC_INFO, SEM_STAT, and SEM_INFO operations are used by the ipcs(1)
283 program to provide information on allocated resources. In the future
284 these may modified or moved to a /proc filesystem interface.
285
286 Various fields in a struct semid_ds were typed as short under Linux 2.2
287 and have become long under Linux 2.4. To take advantage of this, a re‐
288 compilation under glibc-2.1.91 or later should suffice. (The kernel
289 distinguishes old and new calls by an IPC_64 flag in cmd.)
290
291 In some earlier versions of glibc, the semun union was defined in
292 <sys/sem.h>, but POSIX.1 requires that the caller define this union.
293 On versions of glibc where this union is not defined, the macro
294 _SEM_SEMUN_UNDEFINED is defined in <sys/sem.h>.
295
296 The following system limit on semaphore sets affects a semctl() call:
297
298 SEMVMX Maximum value for semval: implementation dependent (32767).
299
300 For greater portability, it is best to always call semctl() with four
301 arguments.
302
303 The sempid value
304 POSIX.1 defines sempid as the "process ID of [the] last operation" on a
305 semaphore, and explicitly notes that this value is set by a successful
306 semop(2) call, with the implication that no other interface affects the
307 sempid value.
308
309 While some implementations conform to the behavior specified in
310 POSIX.1, others do not. (The fault here probably lies with POSIX.1
311 inasmuch as it likely failed to capture the full range of existing im‐
312 plementation behaviors.) Various other implementations also update
313 sempid for the other operations that update the value of a semaphore:
314 the SETVAL and SETALL operations, as well as the semaphore adjustments
315 performed on process termination as a consequence of the use of the
316 SEM_UNDO flag (see semop(2)).
317
318 Linux also updates sempid for SETVAL operations and semaphore adjust‐
319 ments. However, somewhat inconsistently, up to and including Linux
320 4.5, the kernel did not update sempid for SETALL operations. This was
321 rectified in Linux 4.6.
322
324 See shmop(2).
325
327 ipc(2), semget(2), semop(2), capabilities(7), sem_overview(7),
328 sysvipc(7)
329
331 This page is part of release 5.10 of the Linux man-pages project. A
332 description of the project, information about reporting bugs, and the
333 latest version of this page, can be found at
334 https://www.kernel.org/doc/man-pages/.
335
336
337
338Linux 2020-12-21 SEMCTL(2)