1semop(2) System Calls Manual semop(2)
2
6 semop, semtimedop - System V semaphore operations
7
9 Standard C library (libc, -lc)
10
12 #include <sys/sem.h>
13 int semop(int semid, struct sembuf *sops, size_t nsops);
15 int semtimedop(int semid, struct sembuf *sops, size_t nsops,
16 const struct timespec *_Nullable timeout);
17 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
19 semtimedop():
21 _GNU_SOURCE
22
24 Each semaphore in a System V semaphore set has the following associated
25 values:
26 unsigned short semval; /* semaphore value */
28 unsigned short semzcnt; /* # waiting for zero */
29 unsigned short semncnt; /* # waiting for increase */
30 pid_t sempid; /* PID of process that last
31 modified the semaphore value */
32 semop() performs operations on selected semaphores in the set indicated
34 by semid. Each of the nsops elements in the array pointed to by sops
35 is a structure that specifies an operation to be performed on a single
36 semaphore. The elements of this structure are of type struct sembuf,
37 containing the following members:
38 unsigned short sem_num; /* semaphore number */
40 short sem_op; /* semaphore operation */
41 short sem_flg; /* operation flags */
42 Flags recognized in sem_flg are IPC_NOWAIT and SEM_UNDO. If an opera‐
44 tion specifies SEM_UNDO, it will be automatically undone when the
45 process terminates.
46 The set of operations contained in sops is performed in array order,
48 and atomically, that is, the operations are performed either as a com‐
49 plete unit, or not at all. The behavior of the system call if not all
50 operations can be performed immediately depends on the presence of the
51 IPC_NOWAIT flag in the individual sem_flg fields, as noted below.
52 Each operation is performed on the sem_num-th semaphore of the sema‐
54 phore set, where the first semaphore of the set is numbered 0. There
55 are three types of operation, distinguished by the value of sem_op.
56 If sem_op is a positive integer, the operation adds this value to the
58 semaphore value (semval). Furthermore, if SEM_UNDO is specified for
59 this operation, the system subtracts the value sem_op from the sema‐
60 phore adjustment (semadj) value for this semaphore. This operation can
61 always proceed—it never forces a thread to wait. The calling process
62 must have alter permission on the semaphore set.
63 If sem_op is zero, the process must have read permission on the sema‐
65 phore set. This is a "wait-for-zero" operation: if semval is zero, the
66 operation can immediately proceed. Otherwise, if IPC_NOWAIT is speci‐
67 fied in sem_flg, semop() fails with errno set to EAGAIN (and none of
68 the operations in sops is performed). Otherwise, semzcnt (the count of
69 threads waiting until this semaphore's value becomes zero) is incre‐
70 mented by one and the thread sleeps until one of the following occurs:
71 • semval becomes 0, at which time the value of semzcnt is decremented.
73 • The semaphore set is removed: semop() fails, with errno set to EI‐
75 DRM.
76 • The calling thread catches a signal: the value of semzcnt is decre‐
78 mented and semop() fails, with errno set to EINTR.
79 If sem_op is less than zero, the process must have alter permission on
81 the semaphore set. If semval is greater than or equal to the absolute
82 value of sem_op, the operation can proceed immediately: the absolute
83 value of sem_op is subtracted from semval, and, if SEM_UNDO is speci‐
84 fied for this operation, the system adds the absolute value of sem_op
85 to the semaphore adjustment (semadj) value for this semaphore. If the
86 absolute value of sem_op is greater than semval, and IPC_NOWAIT is
87 specified in sem_flg, semop() fails, with errno set to EAGAIN (and none
88 of the operations in sops is performed). Otherwise, semncnt (the
89 counter of threads waiting for this semaphore's value to increase) is
90 incremented by one and the thread sleeps until one of the following oc‐
91 curs:
92 • semval becomes greater than or equal to the absolute value of
94 sem_op: the operation now proceeds, as described above.
95 • The semaphore set is removed from the system: semop() fails, with
97 errno set to EIDRM.
98 • The calling thread catches a signal: the value of semncnt is decre‐
100 mented and semop() fails, with errno set to EINTR.
101 On successful completion, the sempid value for each semaphore specified
103 in the array pointed to by sops is set to the caller's process ID. In
104 addition, the sem_otime is set to the current time.
105 semtimedop()
107 semtimedop() behaves identically to semop() except that in those cases
108 where the calling thread would sleep, the duration of that sleep is
109 limited by the amount of elapsed time specified by the timespec struc‐
110 ture whose address is passed in the timeout argument. (This sleep in‐
111 terval will be rounded up to the system clock granularity, and kernel
112 scheduling delays mean that the interval may overrun by a small
113 amount.) If the specified time limit has been reached, semtimedop()
114 fails with errno set to EAGAIN (and none of the operations in sops is
115 performed). If the timeout argument is NULL, then semtimedop() behaves
116 exactly like semop().
117 Note that if semtimedop() is interrupted by a signal, causing the call
119 to fail with the error EINTR, the contents of timeout are left un‐
120 changed.
121
123 On success, semop() and semtimedop() return 0. On failure, they return
124 -1, and set errno to indicate the error.
125
127 E2BIG The argument nsops is greater than SEMOPM, the maximum number of
128 operations allowed per system call.
129 EACCES The calling process does not have the permissions required to
131 perform the specified semaphore operations, and does not have
132 the CAP_IPC_OWNER capability in the user namespace that governs
133 its IPC namespace.
134 EAGAIN An operation could not proceed immediately and either IPC_NOWAIT
136 was specified in sem_flg or the time limit specified in timeout
137 expired.
138 EFAULT An address specified in either the sops or the timeout argument
140 isn't accessible.
141 EFBIG For some operation the value of sem_num is less than 0 or
143 greater than or equal to the number of semaphores in the set.
144 EIDRM The semaphore set was removed.
146 EINTR While blocked in this system call, the thread caught a signal;
148 see signal(7).
149 EINVAL The semaphore set doesn't exist, or semid is less than zero, or
151 nsops has a nonpositive value.
152 ENOMEM The sem_flg of some operation specified SEM_UNDO and the system
154 does not have enough memory to allocate the undo structure.
155 ERANGE For some operation sem_op+semval is greater than SEMVMX, the im‐
157 plementation dependent maximum value for semval.
158
160 POSIX.1-2008.
161
163 Linux 2.5.52 (backported into Linux 2.4.22), glibc 2.3.3.
164 POSIX.1-2001, SVr4.
165
167 The sem_undo structures of a process aren't inherited by the child pro‐
168 duced by fork(2), but they are inherited across an execve(2) system
169 call.
170 semop() is never automatically restarted after being interrupted by a
172 signal handler, regardless of the setting of the SA_RESTART flag when
173 establishing a signal handler.
174 A semaphore adjustment (semadj) value is a per-process, per-semaphore
176 integer that is the negated sum of all operations performed on a sema‐
177 phore specifying the SEM_UNDO flag. Each process has a list of semadj
178 values—one value for each semaphore on which it has operated using
179 SEM_UNDO. When a process terminates, each of its per-semaphore semadj
180 values is added to the corresponding semaphore, thus undoing the effect
181 of that process's operations on the semaphore (but see BUGS below).
182 When a semaphore's value is directly set using the SETVAL or SETALL re‐
183 quest to semctl(2), the corresponding semadj values in all processes
184 are cleared. The clone(2) CLONE_SYSVSEM flag allows more than one
185 process to share a semadj list; see clone(2) for details.
186 The semval, sempid, semzcnt, and semnct values for a semaphore can all
188 be retrieved using appropriate semctl(2) calls.
189 Semaphore limits
191 The following limits on semaphore set resources affect the semop()
192 call:
193 SEMOPM Maximum number of operations allowed for one semop() call. Be‐
195 fore Linux 3.19, the default value for this limit was 32. Since
196 Linux 3.19, the default value is 500. On Linux, this limit can
197 be read and modified via the third field of /proc/sys/ker‐
198 nel/sem. Note: this limit should not be raised above 1000, be‐
199 cause of the risk of that semop() fails due to kernel memory
200 fragmentation when allocating memory to copy the sops array.
201 SEMVMX Maximum allowable value for semval: implementation dependent
203 (32767).
204 The implementation has no intrinsic limits for the adjust on exit maxi‐
206 mum value (SEMAEM), the system wide maximum number of undo structures
207 (SEMMNU) and the per-process maximum number of undo entries system pa‐
208 rameters.
209
211 When a process terminates, its set of associated semadj structures is
212 used to undo the effect of all of the semaphore operations it performed
213 with the SEM_UNDO flag. This raises a difficulty: if one (or more) of
214 these semaphore adjustments would result in an attempt to decrease a
215 semaphore's value below zero, what should an implementation do? One
216 possible approach would be to block until all the semaphore adjustments
217 could be performed. This is however undesirable since it could force
218 process termination to block for arbitrarily long periods. Another
219 possibility is that such semaphore adjustments could be ignored alto‐
220 gether (somewhat analogously to failing when IPC_NOWAIT is specified
221 for a semaphore operation). Linux adopts a third approach: decreasing
222 the semaphore value as far as possible (i.e., to zero) and allowing
223 process termination to proceed immediately.
224 In Linux 2.6.x, x <= 10, there is a bug that in some circumstances pre‐
226 vents a thread that is waiting for a semaphore value to become zero
227 from being woken up when the value does actually become zero. This bug
228 is fixed in Linux 2.6.11.
229
231 The following code segment uses semop() to atomically wait for the
232 value of semaphore 0 to become zero, and then increment the semaphore
233 value by one.
234 struct sembuf sops[2];
236 int semid;
237 /* Code to set semid omitted */
239 sops[0].sem_num = 0; /* Operate on semaphore 0 */
241 sops[0].sem_op = 0; /* Wait for value to equal 0 */
242 sops[0].sem_flg = 0;
243 sops[1].sem_num = 0; /* Operate on semaphore 0 */
245 sops[1].sem_op = 1; /* Increment value by one */
246 sops[1].sem_flg = 0;
247 if (semop(semid, sops, 2) == -1) {
249 perror("semop");
250 exit(EXIT_FAILURE);
251 }
252 A further example of the use of semop() can be found in shmop(2).
254
256 clone(2), semctl(2), semget(2), sigaction(2), capabilities(7),
257 sem_overview(7), sysvipc(7), time(7)
258Linux man-pages 6.04 2023-03-30 semop(2)