1SEMOP(2) Linux Programmer's Manual SEMOP(2)
2
6 semop, semtimedop - semaphore operations
7
9 #include <sys/types.h>
10 #include <sys/ipc.h>
11 #include <sys/sem.h>
12 int semop(int semid, struct sembuf *sops, unsigned nsops);
14 int semtimedop(int semid, struct sembuf *sops, unsigned nsops, struct
16 timespec *timeout);
17
19 Each semaphore in a semaphore set has the following associated values:
20 unsigned short semval; /* semaphore value */
22 unsigned short semzcnt; /* # waiting for zero */
23 unsigned short semncnt; /* # waiting for increase */
24 pid_t sempid; /* process that did last op */
25 semop() performs operations on selected semaphores in the set indicated
27 by semid. Each of the nsops elements in the array pointed to by sops
28 specifies an operation to be performed on a single semaphore. The ele‐
29 ments of this structure are of type struct sembuf, containing the fol‐
30 lowing members:
31 unsigned short sem_num; /* semaphore number */
33 short sem_op; /* semaphore operation */
34 short sem_flg; /* operation flags */
35 Flags recognized in sem_flg are IPC_NOWAIT and SEM_UNDO. If an opera‐
37 tion specifies SEM_UNDO, it will be automatically undone when the
38 process terminates.
39 The set of operations contained in sops is performed in array order,
41 and atomically, that is, the operations are performed either as a com‐
42 plete unit, or not at all. The behaviour of the system call if not all
43 operations can be performed immediately depends on the presence of the
44 IPC_NOWAIT flag in the individual sem_flg fields, as noted below.
45 Each operation is performed on the sem_num-th semaphore of the sema‐
47 phore set, where the first semaphore of the set is numbered 0. There
48 are three types of operation, distinguished by the value of sem_op.
49 If sem_op is a positive integer, the operation adds this value to the
51 semaphore value (semval). Furthermore, if SEM_UNDO is specified for
52 this operation, the system updates the process undo count (semadj) for
53 this semaphore. This operation can always proceed — it never forces a
54 process to wait. The calling process must have alter permission on the
55 semaphore set.
56 If sem_op is zero, the process must have read permission on the sema‐
58 phore set. This is a "wait-for-zero" operation: if semval is zero, the
59 operation can immediately proceed. Otherwise, if IPC_NOWAIT is speci‐
60 fied in sem_flg, semop() fails with errno set to EAGAIN (and none of
61 the operations in sops is performed). Otherwise semzcnt (the count of
62 processes waiting until this semaphore's value becomes zero) is incre‐
63 mented by one and the process sleeps until one of the following occurs:
64 · semval becomes 0, at which time the value of semzcnt is decre‐
66 mented.
67 · The semaphore set is removed: semop() fails, with errno set to
69 EIDRM.
70 · The calling process catches a signal: the value of semzcnt is
72 decremented and semop() fails, with errno set to EINTR.
73 · The time limit specified by timeout in a semtimedop() call
75 expires: semop() fails, with errno set to EAGAIN.
76 If sem_op is less than zero, the process must have alter permission on
78 the semaphore set. If semval is greater than or equal to the absolute
79 value of sem_op, the operation can proceed immediately: the absolute
80 value of sem_op is subtracted from semval, and, if SEM_UNDO is speci‐
81 fied for this operation, the system updates the process undo count
82 (semadj) for this semaphore. If the absolute value of sem_op is
83 greater than semval, and IPC_NOWAIT is specified in sem_flg, semop()
84 fails, with errno set to EAGAIN (and none of the operations in sops is
85 performed). Otherwise semncnt (the counter of processes waiting for
86 this semaphore's value to increase) is incremented by one and the
87 process sleeps until one of the following occurs:
88 · semval becomes greater than or equal to the absolute value of
90 sem_op, at which time the value of semncnt is decremented, the
91 absolute value of sem_op is subtracted from semval and, if
92 SEM_UNDO is specified for this operation, the system updates the
93 process undo count (semadj) for this semaphore.
94 · The semaphore set is removed from the system: semop() fails,
96 with errno set to EIDRM.
97 · The calling process catches a signal: the value of semncnt is
99 decremented and semop() fails, with errno set to EINTR.
100 · The time limit specified by timeout in a semtimedop() call
102 expires: the system call fails, with errno set to EAGAIN.
103 On successful completion, the sempid value for each semaphore specified
105 in the array pointed to by sops is set to the process ID of the calling
106 process. In addition, the sem_otime is set to the current time.
107 semtimedop() behaves identically to semop() except that in those cases
109 were the calling process would sleep, the duration of that sleep is
110 limited by the amount of elapsed time specified by the timespec struc‐
111 ture whose address is passed in the timeout parameter. If the speci‐
112 fied time limit has been reached, semtimedop() fails with errno set to
113 EAGAIN (and none of the operations in sops is performed). If the time‐
114 out parameter is NULL, then semtimedop() behaves exactly like semop().
115
117 If successful semop() and semtimedop() return 0; otherwise they return
118 -1 with errno indicating the error.
119
121 On failure, errno is set to one of the following:
122 E2BIG The argument nsops is greater than SEMOPM, the maximum number of
124 operations allowed per system call.
125 EACCES The calling process does not have the permissions required to
127 perform the specified semaphore operations, and does not have
128 the CAP_IPC_OWNER capability.
129 EAGAIN An operation could not proceed immediately and either IPC_NOWAIT
131 was specified in sem_flg or the time limit specified in timeout
132 expired.
133 EFAULT An address specified in either the sops or timeout parameters
135 isn't accessible.
136 EFBIG For some operation the value of sem_num is less than 0 or
138 greater than or equal to the number of semaphores in the set.
139 EIDRM The semaphore set was removed.
141 EINTR While blocked in this system call, the process caught a signal.
143 EINVAL The semaphore set doesn't exist, or semid is less than zero, or
145 nsops has a non-positive value.
146 ENOMEM The sem_flg of some operation specified SEM_UNDO and the system
148 does not have enough memory to allocate the undo structure.
149 ERANGE For some operation sem_op+semval is greater than SEMVMX, the
151 implementation dependent maximum value for semval.
152
154 The sem_undo structures of a process aren't inherited across a fork(2)
155 system call, but they are inherited across an execve(2) system call.
156 semop() is never automatically restarted after being interrupted by a
158 signal handler, regardless of the setting of the SA_RESTART flag when
159 establishing a signal handler.
160 semadj is a per-process integer which is simply the (negative) count of
162 all semaphore operations performed specifying the SEM_UNDO flag. When
163 a semaphore's value is directly set using the SETVAL or SETALL request
164 to semctl(2), the corresponding semadj values in all processes are
165 cleared.
166 The semval, sempid, semzcnt, and semnct values for a semaphore can all
168 be retrieved using appropriate semctl(2) calls.
169 The following limits on semaphore set resources affect the semop()
171 call:
172 SEMOPM Maximum number of operations allowed for one semop() call (32)
174 (on Linux, this limit can be read and modified via the third
175 field of /proc/sys/kernel/sem).
176 SEMVMX Maximum allowable value for semval: implementation dependent
178 (32767).
179 The implementation has no intrinsic limits for the adjust on exit maxi‐
181 mum value (SEMAEM), the system wide maximum number of undo structures
182 (SEMMNU) and the per-process maximum number of undo entries system
183 parameters.
184 semtimedop() first appeared in Linux 2.5.52, and was subsequently back‐
186 ported into kernel 2.4.22.
187
189 When a process terminates, its set of associated semadj structures is
190 used to undo the effect of all of the semaphore operations it performed
191 with the SEM_UNDO flag. This raises a difficulty: if one (or more) of
192 these semaphore adjustments would result in an attempt to decrease a
193 semaphore's value below zero, what should an implementation do? One
194 possible approach would be to block until all the semaphore adjustments
195 could be performed. This is however undesirable since it could force
196 process termination to block for arbitrarily long periods. Another
197 possibility is that such semaphore adjustments could be ignored alto‐
198 gether (somewhat analogously to failing when IPC_NOWAIT is specified
199 for a semaphore operation). Linux adopts a third approach: decreasing
200 the semaphore value as far as possible (i.e., to zero) and allowing
201 process termination to proceed immediately.
202 In kernels 2.6.x, x <= 10, there is a bug that in some circumstances
204 prevents a process that is waiting for a semaphore value to become zero
205 from being woken up when the value does actually become zero. This bug
206 is fixed in kernel 2.6.11.
207
209 SVr4, POSIX.1-2001.
210
212 semctl(2), semget(2), sigaction(2), capabilities(7), sem_overview(7),
213 svipc(7)
214Linux 2.6.9 2004-11-10 SEMOP(2)