semop(2)


NAME

   semop, semtimedop - System V semaphore operations

SYNOPSIS

   #include <sys/types.h>
   #include <sys/ipc.h>
   #include <sys/sem.h>

   int semop(int semid, struct sembuf *sops, size_t nsops);

   int semtimedop(int semid, struct sembuf *sops, size_t nsops,
                  const struct timespec *timeout);

   Feature Test Macro Requirements for glibc (see feature_test_macros(7)):

   semtimedop(): _GNU_SOURCE

DESCRIPTION

   Each semaphore in a System V semaphore set has the following associated
   values:

       unsigned short  semval;   /* semaphore value */
       unsigned short  semzcnt;  /* # waiting for zero */
       unsigned short  semncnt;  /* # waiting for increase */
       pid_t           sempid;   /* PID of process that last
                                    modified semaphore value */

   semop() performs operations on selected semaphores in the set indicated
   by  semid.   Each of the nsops elements in the array pointed to by sops
   is a structure that specifies an operation to be performed on a  single
   semaphore.   The  elements of this structure are of type struct sembuf,
   containing the following members:

       unsigned short sem_num;  /* semaphore number */
       short          sem_op;   /* semaphore operation */
       short          sem_flg;  /* operation flags */

   Flags recognized  in  sem_flg  are  IPC_NOWAIT  and  SEM_UNDO.   If  an
   operation  specifies SEM_UNDO, it will be automatically undone when the
   process terminates.

   The set of operations contained in sops is performed  in  array  order,
   and  atomically,  that  is,  the  operations  are performed either as a
   complete unit, or not at all.  The behavior of the system call  if  not
   all  operations can be performed immediately depends on the presence of
   the IPC_NOWAIT flag in the individual sem_flg fields, as noted below.

   Each  operation  is  performed  on  the  sem_num-th  semaphore  of  the
   semaphore  set,  where  the  first  semaphore of the set is numbered 0.
   There are three types of  operation,  distinguished  by  the  value  of
   sem_op.

   If  sem_op  is a positive integer, the operation adds this value to the
   semaphore value (semval).  Furthermore, if SEM_UNDO  is  specified  for
   this  operation,  the  system  subtracts  the  value  sem_op  from  the
   semaphore adjustment (semadj) value for this semaphore.  This operation
   can  always  proceed---it  never  forces  a  thread to wait.  The calling
   process must have alter permission on the semaphore set.

   If sem_op is zero,  the  process  must  have  read  permission  on  the
   semaphore set.  This is a "wait-for-zero" operation: if semval is zero,
   the operation can immediately proceed.   Otherwise,  if  IPC_NOWAIT  is
   specified  in sem_flg, semop() fails with errno set to EAGAIN (and none
   of the operations in sops is performed).  Otherwise, semzcnt (the count
   of  threads  waiting  until  this  semaphore's  value  becomes zero) is
   incremented by one and the thread sleeps until  one  of  the  following
   occurs:

   *  semval becomes 0, at which time the value of semzcnt is decremented.

   *  The  semaphore  set  is  removed:  semop()  fails, with errno set to
      EIDRM.

   *  The calling thread  catches  a  signal:  the  value  of  semzcnt  is
      decremented and semop() fails, with errno set to EINTR.

   If  sem_op is less than zero, the process must have alter permission on
   the semaphore set.  If semval is greater than or equal to the  absolute
   value  of  sem_op,  the operation can proceed immediately: the absolute
   value of  sem_op  is  subtracted  from  semval,  and,  if  SEM_UNDO  is
   specified  for  this  operation,  the system adds the absolute value of
   sem_op to the semaphore adjustment (semadj) value for  this  semaphore.
   If  the absolute value of sem_op is greater than semval, and IPC_NOWAIT
   is specified in sem_flg, semop() fails, with errno set to  EAGAIN  (and
   none  of the operations in sops is performed).  Otherwise, semncnt (the
   counter of threads waiting for this semaphore's value to  increase)  is
   incremented  by  one  and  the thread sleeps until one of the following
   occurs:

   *  semval becomes greater than  or  equal  to  the  absolute  value  of
      sem_op: the operation now proceeds, as described above.

   *  The  semaphore  set  is removed from the system: semop() fails, with
      errno set to EIDRM.

   *  The calling thread  catches  a  signal:  the  value  of  semncnt  is
      decremented and semop() fails, with errno set to EINTR.

   On successful completion, the sempid value for each semaphore specified
   in the array pointed to by sops is set to the caller's process ID.   In
   addition, the sem_otime is set to the current time.

   semtimedop()
   semtimedop()  behaves identically to semop() except that in those cases
   where the calling thread would sleep, the duration  of  that  sleep  is
   limited  by  the  amount  of  elapsed  time  specified  by the timespec
   structure whose address is passed in the timeout argument.  (This sleep
   interval will be rounded up to the system clock granularity, and kernel
   scheduling delays mean  that  the  interval  may  overrun  by  a  small
   amount.)   If  the  specified time limit has been reached, semtimedop()
   fails with errno set to EAGAIN (and none of the operations in  sops  is
   performed).  If the timeout argument is NULL, then semtimedop() behaves
   exactly like semop().

   Note that if semtimedop() is interrupted by a signal, causing the  call
   to  fail  with  the  error  EINTR,  the  contents  of  timeout are left
   unchanged.

RETURN VALUE

   If successful, semop() and semtimedop() return 0; otherwise they return
   -1 with errno indicating the error.

ERRORS

   On failure, errno is set to one of the following:

   E2BIG  The argument nsops is greater than SEMOPM, the maximum number of
          operations allowed per system call.

   EACCES The calling process does not have the  permissions  required  to
          perform  the  specified  semaphore operations, and does not have
          the CAP_IPC_OWNER capability in the user namespace that  governs
          its IPC namespace.

   EAGAIN An operation could not proceed immediately and either IPC_NOWAIT
          was specified in sem_flg or the time limit specified in  timeout
          expired.

   EFAULT An  address specified in either the sops or the timeout argument
          isn't accessible.

   EFBIG  For some operation the value  of  sem_num  is  less  than  0  or
          greater than or equal to the number of semaphores in the set.

   EIDRM  The semaphore set was removed.

   EINTR  While  blocked  in this system call, the thread caught a signal;
          see signal(7).

   EINVAL The semaphore set doesn't exist, or semid is less than zero,  or
          nsops has a nonpositive value.

   ENOMEM The  sem_flg of some operation specified SEM_UNDO and the system
          does not have enough memory to allocate the undo structure.

   ERANGE For some operation sem_op+semval is  greater  than  SEMVMX,  the
          implementation dependent maximum value for semval.

VERSIONS

   semtimedop()  first  appeared  in  Linux  2.5.52,  and was subsequently
   backported into kernel 2.4.22.  Glibc support  for  semtimedop()  first
   appeared in version 2.3.3.

CONFORMING TO

   POSIX.1-2001, POSIX.1-2008, SVr4.

NOTES

   The  inclusion of <sys/types.h> and <sys/ipc.h> isn't required on Linux
   or by any version of POSIX.  However, some old implementations required
   the inclusion of these header files, and the SVID also documented their
   inclusion.  Applications intended to be portable to  such  old  systems
   may need to include these header files.

   The  sem_undo  structures  of  a  process aren't inherited by the child
   produced by fork(2), but they are inherited across an execve(2)  system
   call.

   semop()  is  never automatically restarted after being interrupted by a
   signal handler, regardless of the setting of the SA_RESTART  flag  when
   establishing a signal handler.

   A  semaphore  adjustment (semadj) value is a per-process, per-semaphore
   integer that is the negated  sum  of  all  operations  performed  on  a
   semaphore  specifying  the  SEM_UNDO  flag.  Each process has a list of
   semadj values---one value for each semaphore on  which  it  has  operated
   using  SEM_UNDO.   When a process terminates, each of its per-semaphore
   semadj values is added to the corresponding semaphore, thus undoing the
   effect  of  that  process's  operations  on the semaphore (but see BUGS
   below).  When a semaphore's value is directly set using the  SETVAL  or
   SETALL  request  to  semctl(2),  the corresponding semadj values in all
   processes are cleared.  The clone(2)  CLONE_SYSVSEM  flag  allows  more
   than one process to share a semadj list; see clone(2) for details.

   The  semval, sempid, semzcnt, and semnct values for a semaphore can all
   be retrieved using appropriate semctl(2) calls.

   Semaphore limits
   The following limits on semaphore  set  resources  affect  the  semop()
   call:

   SEMOPM Maximum  number  of  operations  allowed  for  one semop() call.
          Before Linux 3.19, the default value  for  this  limit  was  32.
          Since  Linux  3.19,  the  default  value is 500.  On Linux, this
          limit  can  be  read  and  modified  via  the  third  field   of
          /proc/sys/kernel/sem.   Note:  this  limit  should not be raised
          above 1000, because of the risk of that  semop()  fails  due  to
          kernel  memory  fragmentation when allocating memory to copy the
          sops array.

   SEMVMX Maximum allowable value  for  semval:  implementation  dependent
          (32767).

   The  implementation  has  no  intrinsic  limits  for the adjust on exit
   maximum  value  (SEMAEM),  the  system  wide  maximum  number  of  undo
   structures  (SEMMNU) and the per-process maximum number of undo entries
   system parameters.

BUGS

   When a process terminates, its set of associated semadj  structures  is
   used to undo the effect of all of the semaphore operations it performed
   with the SEM_UNDO flag.  This raises a difficulty: if one (or more)  of
   these  semaphore  adjustments  would result in an attempt to decrease a
   semaphore's value below zero, what should an  implementation  do?   One
   possible approach would be to block until all the semaphore adjustments
   could be performed.  This is however undesirable since it  could  force
   process  termination  to  block  for arbitrarily long periods.  Another
   possibility  is  that  such  semaphore  adjustments  could  be  ignored
   altogether   (somewhat   analogously  to  failing  when  IPC_NOWAIT  is
   specified for a semaphore operation).  Linux adopts a  third  approach:
   decreasing  the  semaphore value as far as possible (i.e., to zero) and
   allowing process termination to proceed immediately.

   In kernels 2.6.x, x <= 10, there is a bug that  in  some  circumstances
   prevents  a thread that is waiting for a semaphore value to become zero
   from being woken up when the value does actually become zero.  This bug
   is fixed in kernel 2.6.11.

EXAMPLE

   The  following  code  segment  uses  semop() to atomically wait for the
   value of semaphore 0 to become zero, and then increment  the  semaphore
   value by one.

       struct sembuf sops[2];
       int semid;

       /* Code to set semid omitted */

       sops[0].sem_num = 0;        /* Operate on semaphore 0 */
       sops[0].sem_op = 0;         /* Wait for value to equal 0 */
       sops[0].sem_flg = 0;

       sops[1].sem_num = 0;        /* Operate on semaphore 0 */
       sops[1].sem_op = 1;         /* Increment value by one */
       sops[1].sem_flg = 0;

       if (semop(semid, sops, 2) == -1) {
           perror("semop");
           exit(EXIT_FAILURE);
       }

SEE ALSO

   clone(2),    semctl(2),   semget(2),   sigaction(2),   capabilities(7),
   sem_overview(7), svipc(7), time(7)

COLOPHON

   This page is part of release 4.09 of the Linux  man-pages  project.   A
   description  of  the project, information about reporting bugs, and the
   latest    version    of    this    page,    can     be     found     at
   https://www.kernel.org/doc/man-pages/.





Opportunity


Personal Opportunity - Free software gives you access to billions of dollars of software at no cost. Use this software for your business, personal use or to develop a profitable skill. Access to source code provides access to a level of capabilities/information that companies protect though copyrights. Open source is a core component of the Internet and it is available to you. Leverage the billions of dollars in resources and capabilities to build a career, establish a business or change the world. The potential is endless for those who understand the opportunity.

Business Opportunity - Goldman Sachs, IBM and countless large corporations are leveraging open source to reduce costs, develop products and increase their bottom lines. Learn what these companies know about open source and how open source can give you the advantage.





Free Software


Free Software provides computer programs and capabilities at no cost but more importantly, it provides the freedom to run, edit, contribute to, and share the software. The importance of free software is a matter of access, not price. Software at no cost is a benefit but ownership rights to the software and source code is far more significant.


Free Office Software - The Libre Office suite provides top desktop productivity tools for free. This includes, a word processor, spreadsheet, presentation engine, drawing and flowcharting, database and math applications. Libre Office is available for Linux or Windows.





Free Books


The Free Books Library is a collection of thousands of the most popular public domain books in an online readable format. The collection includes great classical literature and more recent works where the U.S. copyright has expired. These books are yours to read and use without restrictions.


Source Code - Want to change a program or know how it works? Open Source provides the source code for its programs so that anyone can use, modify or learn how to write those programs themselves. Visit the GNU source code repositories to download the source.





Education


Study at Harvard, Stanford or MIT - Open edX provides free online courses from Harvard, MIT, Columbia, UC Berkeley and other top Universities. Hundreds of courses for almost all major subjects and course levels. Open edx also offers some paid courses and selected certifications.


Linux Manual Pages - A man or manual page is a form of software documentation found on Linux/Unix operating systems. Topics covered include computer programs (including library and system calls), formal standards and conventions, and even abstract concepts.