[RFC] locking.9

From: Julian Elischer <julian_at_elischer.org>
Date: Tue, 13 Mar 2007 16:59:35 -0700
resend to the right place with a corrected attachment.

Robert Watson wrote:
> 
> The idea of a locking(9) has been kicked around for a while, and its 
> time has definitely come.  It would summarize the properties and cross 
> reference the man pages of the various primitives, and suggest a 
> preference and strategy for using them in new code vs. existing code, 
> etc.  Distinguishing dimensions would include things like whether it is 
> sleepable, supports priority propagation, can be acquired in interrupt 
> context (a result of the prior two properties), whether it is fair, etc, 
> etc.  And include stern warnings about not using lockmgr in new code :-).
> 
> Robert N M Watson
> Computer Laboratory
> University of Cambridge


ok so how about I commit this to get us started and the nroff and
locking experts can take it from there.



.\"
.\" Copyright (c) 1998 Berkeley Software Design, Inc. 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.
.\" 3. Berkeley Software Design Inc's name may not be used to endorse or
.\"    promote products derived from this software without specific prior
.\"    written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY BERKELEY SOFTWARE DESIGN INC ``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 BERKELEY SOFTWARE DESIGN INC 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.
.\"
.\"	from BSDI $Id: mutex.4,v 1.1.2.3 1998/04/27 22:53:13 ewv Exp $
.\" $FreeBSD: src/share/man/man9/mutex.9,v 1.54 2007/03/09 22:41:01 jhb Exp $
.\"
.Dd March 14, 2007
.Dt LOCKING 9
.Os
.Sh NAME
.Nm locking ,
.Nd kernel synchronization primitives
.Sh SYNOPSIS
All sorts of stuff to go here.
.Pp
.Sh DESCRIPTION
The 
.Em
FreeBSD
kernel is written to run across multiple CPUs and as such requires
several different sychronization primatives to allow the developers 
to safely access and manipulate the many data types required.
.Pp
These include:
.Bl -enum
.It
Spin Mutexes
.It
Sleep Mutexes
.It
pool Mutexes
.It
Shared-Exclusive locks
.It
Reader-Writer locks
.It
Turnstyles
.It
Semaphores
.It
Condition variables
.It
Sleep/wakeup
.It
Giant
.It
Lockmanager locks
.El

.Pp
The primnatives interact and have a number of rules regarding how
they can and can not be combined. Ther eare too many for the average humanmind and they
keep changing. (if you disagree, please write replacement text)  :-)
.Pp
Some of these primatives may be used at the low (interrupt) level and some may not.
.Pp
There are strict ordering requirements and for some of the types this
is checked using the 
.Xr witness 4
code.
.Pp
.Ss SPIN Mutexes
Mutexes are the basic primative.
You either hold it or you don't.
If you don't own it then you just spin, waiting for teh holder (on another CPU)
to release it. Hopefully they are doing something fast. You can not do anythign that 
deschedules the thread while you are holding a SPIN mutex.
.Ss Sleep Mutexes
Basically sleep (regular) mutexes will deschedule the thread if
the mutex can not be acquired. As in spin mutexes, you either get it or you don't.
You may call the
.Xr sleep 9
call
.Fn msleep
or the new 
.Fn mtx_sleep
variant. These will atomically drop the mutex and reacquire it
as part of waking up.
.Ss Pool Mutexes
A variant of SLEEP mutexes where the allocation of the mutex is handled more by the system.
.Ss Sx_locks
Shared/exclusive locks are used to protect data that are read far more often
than they are written.
Mutexes are inherently more efficient than shared/exclusive locks, so
shared/exclusive locks should be used prudently.
A thread may hold a shared or exclusive lock on an
.Em sx_lock
lock while sleeping.
As a result, an
.Em sx_lockm
lock may not be acquired while holding a mutex.
Otherwise, if one thread slept while holding an
.Em sx_lockm
lock while another thread blocked on the same
.Em sx_lockm
lock after acquiring a mutex, then the second thread would effectively
end up sleeping while holding a mutex, which is not allowed.
.Ss Rw_locks
Reader/writer locks allow shared access to protected data by multiple threads,
or exclusive access by a single thread.
The threads with shared access are known as
.Em readers
since they only read the protected data.
A thread with exclusive access is known as a
.Em writer
since it can modify protected data.
.Pp
Although reader/writer locks look very similar to
.Xr sx 9
locks, their usage pattern is different.
Reader/writer locks can be treated as mutexes (see
.Xr mutex 9 )
with shared/exclusive semantics.
Unlike
.Xr sx 9 ,
an
.Em rw_lock
can be locked while holding a non-spin mutex, and an
.Em rw_lock
cannot be held while sleeping.
The
.Em rw_lock
locks have priority propagation like mutexes, but priority
can be propagated only to an exclusive holder.
This limitation comes from the fact that shared owners
are anonymous.
Another important property is that shared holders of
.Em rw_lock
can recurse,
but exclusive locks are not allowed to recurse.
.Ss Turnstyless
Turnstiles are used to hold a queue of threads blocked on
non-sleepable locks.  Sleepable locks use condition variables to
implement their queues.  Turnstiles differ from a sleep queue in that
turnstile queue's are assigned to a lock held by an owning thread.  Thus,
when one thread is enqueued onto a turnstile, it can lend its priority
to the owning thread.
.Ss Semaphores
.Ss Condition variables
Condition variables are used in conjunction with mutexes to wait for conditions
to occur.
A thread must hold the mutex before calling the
.Fn cv_wait* ,
functions.
When a thread waits on a condition, the mutex
is atomically released before the thread is blocked, then reacquired
before the function call returns.
.Ss Giant
Giant is a special instance of a sleep lock.
it has several special characteristics.
.Ss Sleep/wakeup
The functions
.Fn tsleep ,
.Fn msleep ,
.Fn msleep_spin ,
.Fn pause ,
.Fn wakeup ,
and
.Fn wakeup_one
handle event-based thread blocking.
If a thread must wait for an
external event, it is put to sleep by
.Fn tsleep ,
.Fn msleep ,
.Fn msleep_spin ,
or
.Fn pause .
Threads may also wait using one of the locking primitive sleep routines
.Xr mtx_sleep 9 ,
.Xr rw_sleep 9 ,
or
.Xr sx_sleep 9 .
.Pp
The parameter
.Fa chan
is an arbitrary address that uniquely identifies the event on which
the thread is being put to sleep.
All threads sleeping on a single
.Fa chan
are woken up later by
.Fn wakeup ,
often called from inside an interrupt routine, to indicate that the
resource the thread was blocking on is available now.
.Pp
Several of the sleep functions including
.Fn msleep ,
.Fn msleep_spin ,
and the locking primitive sleep routines specify an additional lock
parameter.
The lock will be released before sleeping and reacquired
before the sleep routine returns.
If
.Fa priority
includes the
.Dv PDROP
flag, then
the lock will not be reacquired before returning.
The lock is used to ensure that a condition can be checked atomically,
and that the current thread can be suspended without missing a
change to the condition, or an associated wakeup.
In addition, all of the sleep routines will fully drop the
.Va Giant
mutex
(even if recursed)
while the thread is suspended and will reacquire the
.Va Giant
mutex before the function returns.
.Pp
.Ss lockmanager locks
Largly deprecated. See the 
.Xr lock 9
page for more information.
I don't know what the downsides are but I'm sure someone will fill in this part.
.Sh Interaction tables.
the following table shows what you can an d can not do if you hold
one of the synchronisation primatives discussed here:
(someone who knows what they are talking about should write this table)
.Bl -column ".Ic You have:  You want:" ".Xr sleep" ".Xr sc_lock" ".Xr rw_lock" -offset indent
.It Xo
.Em "You have:  You want:" Ta Xr sleep Ta Xr sx_lock Ta Xr rw_lock
.Xc
.It Ic SPIN mutex Ta \&tough Ta \&tough Ta \&tough
.It Ic Sleep mutex Ta \&ok Ta \&??? Ta \&???
.El
.Sh SEE ALSO
.Xr condvar 9 ,
.Xr lock 9
.Xr mtx_pool 9 ,
.Xr rwlock 9 ,
.Xr sema 9 ,
.Xr sleep 9 ,
.Xr sx 9
.Xr LOCK_PROFILING 9 ,
.Xr WITNESS 9 ,
.Sh HISTORY
These
functions appeared in
.Bsx 4.1 
through
.Fx 7.0 
Received on Tue Mar 13 2007 - 23:08:03 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:06 UTC