xref: /freebsd/share/man/man9/rmlock.9 (revision 7750ad47a9a7dbc83f87158464170c8640723293)
1.\" Copyright (c) 2007 Stephan Uphoff <ups@FreeBSD.org>
2.\" Copyright (c) 2006 Gleb Smirnoff <glebius@FreeBSD.org>
3.\" All rights reserved.
4.\"
5.\" Redistribution and use in source and binary forms, with or without
6.\" modification, are permitted provided that the following conditions
7.\" are met:
8.\" 1. Redistributions of source code must retain the above copyright
9.\"    notice, this list of conditions and the following disclaimer.
10.\" 2. Redistributions in binary form must reproduce the above copyright
11.\"    notice, this list of conditions and the following disclaimer in the
12.\"    documentation and/or other materials provided with the distribution.
13.\"
14.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24.\" SUCH DAMAGE.
25.\"
26.\" $FreeBSD$
27.\"
28.\" Based on rwlock.9 man page
29.Dd November 16, 2011
30.Dt RMLOCK 9
31.Os
32.Sh NAME
33.Nm rmlock ,
34.Nm rm_init ,
35.Nm rm_init_flags ,
36.Nm rm_destroy ,
37.Nm rm_rlock ,
38.Nm rm_try_rlock ,
39.Nm rm_wlock ,
40.Nm rm_runlock ,
41.Nm rm_wunlock ,
42.Nm rm_wowned ,
43.Nm RM_SYSINIT
44.Nd kernel reader/writer lock optimized for mostly read access patterns
45.Sh SYNOPSIS
46.In sys/param.h
47.In sys/lock.h
48.In sys/rmlock.h
49.Ft void
50.Fn rm_init "struct rmlock *rm" "const char *name"
51.Ft void
52.Fn rm_init_flags "struct rmlock *rm" "const char *name" "int opts"
53.Ft void
54.Fn rm_destroy "struct rmlock *rm"
55.Ft void
56.Fn rm_rlock "struct rmlock *rm"  "struct rm_priotracker* tracker"
57.Ft int
58.Fn rm_try_rlock "struct rmlock *rm"  "struct rm_priotracker* tracker"
59.Ft void
60.Fn rm_wlock "struct rmlock *rm"
61.Ft void
62.Fn rm_runlock "struct rmlock *rm" "struct rm_priotracker* tracker"
63.Ft void
64.Fn rm_wunlock "struct rmlock *rm"
65.Ft int
66.Fn rm_wowned "const struct rmlock *rm"
67.In sys/kernel.h
68.Fn RM_SYSINIT "name" "struct rmlock *rm" "const char *desc" "int opts"
69.Sh DESCRIPTION
70Mostly reader locks allow shared access to protected data by multiple threads,
71or exclusive access by a single thread.
72The threads with shared access are known as
73.Em readers
74since they only read the protected data.
75A thread with exclusive access is known as a
76.Em writer
77since it can modify protected data.
78.Pp
79Read mostly locks are designed to be efficient for locks almost exclusively
80used as reader locks and as such should be used for protecting data that
81rarely changes.
82Acquiring an exclusive lock after the lock had been locked for shared access
83is an expensive operation.
84.Pp
85Although reader/writer locks look very similar to
86.Xr sx 9
87locks, their usage pattern is different.
88Reader/writer locks can be treated as mutexes (see
89.Xr mutex 9 )
90with shared/exclusive semantics unless initialized with
91.Dv RM_SLEEPABLE .
92Unlike
93.Xr sx 9 ,
94an
95.Nm
96can be locked while holding a non-spin mutex, and an
97.Nm
98cannot be held while sleeping, again unless initialized with
99.Dv RM_SLEEPABLE .
100The
101.Nm
102locks have full priority propagation like mutexes.
103The
104.Va rm_priotracker
105structure argument supplied in
106.Fn rm_rlock
107and
108.Fn rm_runlock
109is used to keep track of the read owner(s).
110Another important property is that shared holders of
111.Nm
112can recurse if the lock has been initialized with the
113.Dv LO_RECURSABLE
114option, however exclusive locks are not allowed to recurse.
115.Ss Macros and Functions
116.Bl -tag -width indent
117.It Fn rm_init "struct rmlock *rm" "const char *name"
118Initialize structure located at
119.Fa rm
120as mostly reader lock, described by
121.Fa name .
122The name description is used solely for debugging purposes.
123This function must be called before any other operations
124on the lock.
125.It Fn rm_init_flags "struct rmlock *rm" "const char *name" "int opts"
126Initialize the rm lock just like the
127.Fn rm_init
128function, but specifying a set of optional flags to alter the
129behaviour of
130.Fa rm ,
131through the
132.Fa opts
133argument.
134It contains one or more of the following flags:
135.Bl -tag -width ".Dv RM_NOWITNESS"
136.It Dv RM_NOWITNESS
137Instruct
138.Xr witness 4
139to ignore this lock.
140.It Dv RM_RECURSE
141Allow threads to recursively acquire exclusive locks for
142.Fa rm .
143.It Dv RM_SLEEPABLE
144Allow writers to sleep while holding the lock.
145Readers must not sleep while holding the lock and can avoid to sleep on
146taking the lock by using
147.Fn rm_try_rlock
148instead of
149.Fn rm_rlock .
150.El
151.It Fn rm_rlock "struct rmlock *rm" "struct rm_priotracker* tracker"
152Lock
153.Fa rm
154as a reader.
155Using
156.Fa tracker
157to track read owners of a lock for priority propagation.
158This data structure is only used internally by
159.Nm
160and must persist until
161.Fn rm_runlock
162has been called.
163This data structure can be allocated on the stack since
164rmlocks cannot be held while sleeping.
165If any thread holds this lock exclusively, the current thread blocks,
166and its priority is propagated to the exclusive holder.
167If the lock was initialized with the
168.Dv LO_RECURSABLE
169option the
170.Fn rm_rlock
171function can be called when the thread has already acquired reader
172access on
173.Fa rm .
174This is called
175.Dq "recursing on a lock" .
176.It Fn rm_try_rlock "struct rmlock *rm" "struct rm_priotracker* tracker"
177Try to lock
178.Fa rm
179as a reader.
180.Fn rm_try_rlock
181will return 0 if the lock cannot be acquired immediately;
182otherwise the lock will be acquired and a non-zero value will be returned.
183Note that
184.Fn rm_try_rlock
185may fail even while the lock is not currently held by a writer.
186.It Fn rm_wlock "struct rmlock *rm"
187Lock
188.Fa rm
189as a writer.
190If there are any shared owners of the lock, the current thread blocks.
191The
192.Fn rm_wlock
193function cannot be called recursively.
194.It Fn rm_runlock "struct rmlock *rm" "struct rm_priotracker* tracker"
195This function releases a shared lock previously acquired by
196.Fn rm_rlock .
197The
198.Fa tracker
199argument must match the
200.Fa tracker
201argument used for acquiring the shared lock
202.It Fn rm_wunlock "struct rmlock *rm"
203This function releases an exclusive lock previously acquired by
204.Fn rm_wlock .
205.It Fn rm_destroy "struct rmlock *rm"
206This functions destroys a lock previously initialized with
207.Fn rm_init .
208The
209.Fa rm
210lock must be unlocked.
211.It Fn rm_wowned "const struct rmlock *rm"
212This function returns a non-zero value if the current thread owns an
213exclusive lock on
214.Fa rm .
215.El
216.Sh SEE ALSO
217.Xr locking 9 ,
218.Xr mutex 9 ,
219.Xr panic 9 ,
220.Xr rwlock 9 ,
221.Xr sema 9 ,
222.Xr sx 9
223.Sh HISTORY
224These
225functions appeared in
226.Fx 7.0 .
227.Sh AUTHORS
228.An -nosplit
229The
230.Nm
231facility was written by
232.An "Stephan Uphoff" .
233This manual page was written by
234.An "Gleb Smirnoff"
235for rwlock and modified to reflect rmlock by
236.An "Stephan Uphoff" .
237.Sh BUGS
238The
239.Nm
240implementation is currently not optimized for single processor systems.
241.Pp
242.Fn rm_try_rlock
243can fail transiently even when there is no writer, while another reader
244updates the state on the local CPU.
245.Pp
246The
247.Nm
248implementation uses a single per CPU list shared by all
249rmlocks in the system.
250If rmlocks become popular, hashing to multiple per CPU queues may
251be needed to speed up the writer lock process.
252.Pp
253The
254.Nm
255can currently not be used as a lock argument for condition variable
256wait functions.
257