1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-28 11:57:28 +00:00
freebsd/lib/libc/sys/semop.2
Michael Reifenberger 91a701cd13 Fix SysV Semaphore Handling.
Updated by peter following KSE and Giant pushdown.
I've running with this patch for two week with no ill side effects.

PR:		kern/12014: Fix SysV Semaphore handling
Submitted by:	Peter Jeremy <peter.jeremy@alcatel.com.au>
2001-10-11 08:15:14 +00:00

275 lines
6.9 KiB
Groff

.\"
.\" Copyright (c) 1995 David Hovemeyer <daveho@infocom.com>
.\"
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE DEVELOPERS ``AS IS'' AND ANY EXPRESS OR
.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
.\" IN NO EVENT SHALL THE DEVELOPERS BE LIABLE FOR ANY DIRECT, INDIRECT,
.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd September 22, 1995
.Dt SEMOP 2
.Os
.Sh NAME
.Nm semop
.Nd atomic array of operations on a semaphore set
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
.In sys/types.h
.In sys/ipc.h
.In sys/sem.h
.Ft int
.Fn semop "int semid" "struct sembuf array[]" "unsigned nops"
.Sh DESCRIPTION
.Fn Semop
atomically performs the array of operations indicated by
.Fa array
on the semaphore set indicated by
.Fa semid .
The length of
.Fa array
is indicated by
.Fa nops .
Each operation is encoded in a
.Fa "struct sembuf" ,
which is defined as follows:
.Bd -literal
.\"
.\" From <sys/sem.h>
.\"
struct sembuf {
u_short sem_num; /* semaphore # */
short sem_op; /* semaphore operation */
short sem_flg; /* operation flags */
};
.Ed
.Pp
For each element in
.Fa array ,
.Fa sem_op
and
.Fa sem_flg
determine an operation to be performed on semaphore number
.Fa sem_num
in the set. The values
.Dv SEM_UNDO
and
.Dv IPC_NOWAIT
may be
.Em OR Ns 'ed
into the
.Fa sem_flg
member in order to modify the behavior of the given operation.
.Pp
The operation performed depends as follows on the value of
.Fa sem_op :
.\"
.\" This section is based on the description of semop() in
.\" Stevens, _Advanced Programming in the UNIX Environment_,
.\" and the semop(2) description in The Open Group Unix2 specification.
.\"
.Bl -bullet
.It
When
.Fa sem_op
is positive and the process has alter permission,
the semaphore's value is incremented by
.Fa sem_op Ns 's
value. If
.Dv SEM_UNDO
is specified, the semaphore's adjust on exit value is decremented by
.Fa sem_op Ns 's
value. A positive value for
.Fa sem_op
generally corresponds to a process releasing a resource
associated with the semaphore.
.It
The behavior when
.Fa sem_op
is negative and the process has alter permission,
depends on the current value of the semaphore:
.Bl -bullet
.It
If the current value of the semaphore is greater than or equal to
the absolute value of
.Fa sem_op ,
then the value is decremented by the absolute value of
.Fa sem_op .
If
.Dv SEM_UNDO
is specified, the semaphore's adjust on exit
value is incremented by the absolute value of
.Fa sem_op .
.It
If the current value of the semaphore is less than the absolute value of
.Fa sem_op ,
one of the following happens:
.\" XXX a *second* sublist?
.Bl -bullet
.It
If
.Dv IPC_NOWAIT
was specified, then
.Fn semop
returns immediately with a return value of
.Er EAGAIN .
.It
Otherwise, the calling process is put to sleep until one of the following
conditions is satisfied:
.\" XXX We already have two sublists, why not a third?
.Bl -bullet
.It
Some other process removes the semaphore with the
.Dv IPC_RMID
option of
.Fn semctl .
In this case,
.Fn semop
returns immediately with a return value of
.Er EIDRM .
.It
The process receives a signal that is to be caught.
In this case, the process will resume execution as defined by
.Fn sigaction .
.It
The semaphore's
value is greater than or equal to the absolute value of
.Fa sem_op .
When this condition becomes true, the semaphore's value is decremented
by the absolute value of
.Fa sem_op ,
the semaphore's adjust on exit value is incremented by the
absolute value of
.Fa sem_op .
.El
.El
.Pp
A negative value for
.Fa sem_op
generally means that a process is waiting for a resource to become
available.
.El
.Pp
.It
When
.Fa sem_op
is zero and the process has read permission,
one of the following will occur:
.Bl -bullet
.It
If the current value of the semaphore is equal to zero
then
.Fn semop
can return immediately.
.It
If
.Dv IPC_NOWAIT
was specified, then
.Fn semop
returns immediately with a return value of
.Er EAGAIN .
.It
Otherwise, the calling process is put to sleep until one of the following
conditions is satisfied:
.\" XXX Another nested sublists
.Bl -bullet
.It
Some other process removes the semaphore with the
.Dv IPC_RMID
option of
.Fn semctl .
In this case,
.Fn semop
returns immediately with a return value of
.Er EIDRM .
.It
The process receives a signal that is to be caught.
In this case, the process will resume execution as defined by
.Fn sigaction
.It
The semaphore's value becomes zero.
.El
.El
.Pp
For each semaphore a process has in use, the kernel maintains an
`adjust on exit' value, as alluded to earlier. When a process
exits, either voluntarily or involuntarily, the adjust on exit value
for each semaphore is added to the semaphore's value. This can
be used to insure that a resource is released if a process terminates
unexpectedly.
.Sh RETURN VALUES
.Rv -std semop
.Sh ERRORS
.Fn Semop
will fail if:
.Bl -tag -width Er
.It Bq Er EINVAL
No semaphore set corresponds to
.Fa semid ,
or the process would exceed the system-defined limit for the number of
per-process SEM_UNDO structures.
.It Bq Er EACCES
Permission denied due to mismatch between operation and mode of
semaphore set.
.It Bq Er EAGAIN
The semaphore's value would have resulted in the process being put to sleep
and
.Dv IPC_NOWAIT
was specified.
.It Bq Er E2BIG
Too many operations were specified.
.Bq Dv SEMOPM
.It Bq Er EFBIG
.\"
.\" I'd have thought this would be EINVAL, but the source says
.\" EFBIG.
.\"
.Fa sem_num
was not in the range of valid semaphores for the set.
.It Bq Er EIDRM
The semaphore set was removed from the system.
.It Bq Er EINTR
The
.Fn semop
call was interrupted by a signal.
.It Bq Er ENOSPC
The system SEM_UNDO pool
.Bq Dv SEMMNU
is full.
.It Bq Er ERANGE
The requested operation would cause either
the semaphore's current value
.Bq Dv SEMVMX
or its adjust on exit value
.Bq Dv SEMAEM
to exceed the system-imposed limits.
.El
.Sh SEE ALSO
.Xr semctl 2 ,
.Xr semget 2 ,
.Xr sigaction 2
.Sh BUGS
.Fn Semop
may block waiting for memory even if
.Dv IPC_NOWAIT
is specified.