NetBSD/share/man/man9/mutex.9

274 lines
8.4 KiB
Groff

.\" $NetBSD: mutex.9,v 1.9 2007/04/09 13:37:57 ad Exp $
.\"
.\" Copyright (c) 2007 The NetBSD Foundation, Inc.
.\" All rights reserved.
.\"
.\" This code is derived from software contributed to The NetBSD Foundation
.\" by Andrew Doran.
.\"
.\" 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.
.\" 3. All advertising materials mentioning features or use of this software
.\" must display the following acknowledgement:
.\" This product includes software developed by the NetBSD
.\" Foundation, Inc. and its contributors.
.\" 4. Neither the name of The NetBSD Foundation nor the names of its
.\" contributors may be used to endorse or promote products derived
.\" from this software without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
.\" ``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 FOUNDATION OR CONTRIBUTORS
.\" 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.
.\"
.Dd April 9, 2007
.Dt MUTEX 9
.Os
.Sh NAME
.Nm mutex ,
.Nm mutex_init ,
.Nm mutex_destroy ,
.Nm mutex_enter ,
.Nm mutex_exit ,
.Nm mutex_tryenter ,
.Nm mutex_owned
.Nd mutual exclusion primitives
.Sh SYNOPSIS
.In sys/mutex.h
.Ft void
.Fn mutex_init "kmutex_t *mtx" "kmutex_type_t type" "int ipl"
.Ft void
.Fn mutex_destroy "kmutex_t *mtx"
.Ft void
.Fn mutex_enter "kmutex_t *mtx"
.Ft void
.Fn mutex_exit "kmutex_t *mtx"
.Ft int
.Fn mutex_tryenter "kmutex_t *mtx"
.Ft int
.Fn mutex_owned "kmutex_t *mtx"
.Pp
.Cd "options DIAGNOSTIC"
.Cd "options LOCKDEBUG"
.Sh DESCRIPTION
Mutexes are used in the kernel to implement mutual exclusion among LWPs
(lightweight processes) and interrupt handlers.
.Pp
The
.Vt kmutex_t
type provides storage for the mutex object.
This should be treated as an opaque object and not examined directly by
consumers.
.Pp
Mutexes replace the
.Xr spl 9
system traditionally used to provide synchronization between interrupt
handlers and LWPs, and in combination with reader / writer locks replace
the
.Xr lockmgr 9
facility.
.Sh OPTIONS
.Bl -tag -width abcd
.It Cd "options DIAGNOSTIC"
.Pp
Kernels compiled with the
.Dv DIAGNOSTIC
option perform basic sanity checks on mutex operations.
.It Cd "options LOCKDEBUG"
.Pp
Kernels compiled with the
.Dv LOCKDEBUG
option perform potentially CPU intensive sanity checks
on mutex operations.
.El
.Sh FUNCTIONS
.Bl -tag -width abcd
.It Fn mutex_init "mtx" "type" "ipl"
.Pp
Dynamically initialize a mutex for use.
No other operations can be performed on a mutex until it has been initialized.
.Fn mutex_init
may block in order to allocate memory.
.Pp
The
.Fa type
argument specifies the kind of mutex to be initialized, controlling the
behaviour of the mutex.
Once initialized, all types of mutex are manipulated using the same interface.
Valid types are as follows:
.Bl -tag -width cdoscdosrunru
.It Dv MUTEX_DEFAULT
Requests an adaptive mutex.
Adaptive mutexes provide mutual exclusion between LWPs.
.Pp
When initializing an adaptive mutex,
.Dv IPL_NONE
must be specified as the
.Fa ipl
argument.
Adaptive mutexes should not be acquired from an interrupt handler.
.Pp
An LWP may either sleep or busy-wait when attempting to acquire
an adaptive mutex that is already held.
.It Dv MUTEX_SPIN
Requests a spin mutex.
Spin mutexes provide mutual exclusion between LWPs, and between LWPs
and interrupt handlers.
.Pp
When initializing a spin mutex, the
.Fa ipl
argument is used to pass an system interrupt priority level (SPL)
that will block all interrupt handlers that may try to acquire the mutex.
.Pp
LWPs that own spin mutexes may not sleep, and therefore must not
try to acquire adaptive mutexes or other sleep locks.
.Pp
A processor will always busy-wait when attempting to acquire
a spin mutex that is already held.
.El
.Pp
In general, device drivers should not directly request spin or adaptive
mutexes.
For device drivers a third type of mutex is provided:
.Bl -tag -width cdoscdosrunru
.It Dv MUTEX_DRIVER
Requests a device driver mutex.
.Pp
The
.Fa ipl
argument is used to determine whether a spin or adaptive mutex is
returned, depending on how interrupt handling is implemented by
the machine architecture.
.Pp
If
.Dv IPL_NONE
is specified,
.Fn mutex_init
is guaranteed to return an adaptive mutex that can be used to
provide mutual exclusion for device driver code that does not
run from an interrupt handler.
Other levels return mutexes that can be used to synchronize
with interrupt handlers.
For example, to request a mutex for synchronizing with network
interrupt handlers, specify
.Dv IPL_NET .
Beyond this, device drivers should not make assumptions about
the type of mutex returned.
.El
.It Fn mutex_destroy "mtx"
.Pp
Release resources used by a mutex.
The mutex may not be used after it has been destroyed.
.Fn mutex_destroy
may block in order to free memory.
.It Fn mutex_enter "mtx"
.Pp
Acquire a mutex.
If the mutex is already held, the caller will block and not return until the
mutex is acquired.
.Pp
Mutexes and other types of locks must always be acquired in a
consistent order with respect to each other.
Otherwise, the potential for system deadlock exists.
.Pp
Adaptive mutexes and other types of lock that can sleep may
not be acquired once a spin mutex is held by the caller.
Note that device driver mutexes at a level other than
.Dv IPL_NONE
can be spin mutexes.
.It Fn mutex_exit "mtx"
.Pp
Release a mutex.
The mutex must have been previously acquired by the caller.
Mutexes may be released out of order as needed.
.It Fn mutex_tryenter "mtx"
.Pp
Try to acquire a mutex, but do not block if the mutex is already held.
Returns non-zero if the mutex was acquired, or zero if the mutex was
already held.
.Pp
.Fn mutex_tryenter
can be used as an optimization when acquiring locks in the the wrong order.
For example, in a setting where the convention is that
.Dv first_lock
must be acquired before
.Dv second_lock ,
the following can be used to optimistically lock in reverse order:
.Bd -literal
/* We hold second_lock, but not first_lock. */
KASSERT(mutex_owned(\*[Am]second_lock));
if (!mutex_tryenter(\*[Am]first_lock)) {
/* Failed to get it - lock in the correct order. */
mutex_exit(\*[Am]second_lock);
mutex_enter(\*[Am]first_lock);
mutex_enter(\*[Am]second_lock);
/*
* We may need to recheck any conditions the code
* path depends on, as we released second_lock
* briefly.
*/
}
.Ed
.It Fn mutex_owned "mtx"
.Pp
For adaptive mutexes, return non-zero if the current LWP holds the mutex.
For spin mutexes, return non-zero if the mutex is held, potentially by the
current processor.
Otherwise, return zero.
.Pp
.Fn mutex_owned
is provided for making diagnostic checks to verify that a lock is held.
For example:
.Bd -literal
KASSERT(mutex_owned(\*[Am]driver_lock));
.Ed
.Pp
It should not be used to make locking decisions at run time, or to
verify that a lock is unheld.
.El
.Sh CODE REFERENCES
This section describes places within the
.Nx
source tree where code implementing mutexes can be found.
All pathnames are relative to
.Pa /usr/src .
.Pp
The core of the mutex implementation is in
.Pa sys/kern/kern_mutex.c .
.Pp
The header file
.Pa sys/sys/mutex.h
describes the public interface, and interfaces that machine-dependent
code must provide to support mutexes.
.Sh SEE ALSO
.Xr condvar 9 ,
.Xr mb 9 ,
.Xr rwlock 9
.Pp
.Rs
.%A Jim Mauro
.%A Richard McDougall
.%T Solaris Internals: Core Kernel Architecture ,
.%I Prentice Hall
.%D 2001
.%O ISBN 0-13-022496-0
.Re
.Sh HISTORY
The mutex primitives first appeared in
.Nx 5.0 .