1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
|
.\" $NetBSD: __cpu_simple_lock.9,v 1.2 2024/10/26 03:05:06 riastradh Exp $
.\"
.\" Copyright (c) 2022 The NetBSD Foundation, 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.
.\"
.\" 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 October 25, 2024
.Dt __CPU_SIMPLE_LOCK 9
.Os
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh NAME
.Nm __cpu_simple_lock
.Nd simple spin locks
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh SYNOPSIS
.In sys/lock.h
.\"
.Ft void
.Fn __cpu_simple_lock_init "__cpu_simple_lock_t *lock"
.\"
.Vt #define __SIMPLELOCK_UNLOCKED ...
.\"
.Ft void
.Fn __cpu_simple_lock "__cpu_simple_lock_t *lock"
.Ft int
.Fn __cpu_simple_lock_try "__cpu_simple_lock *lock"
.Ft void
.Fn __cpu_simple_unlock "__cpu_simple_lock_t *lock"
.\"
.Ft int
.Fn __SIMPLELOCK_LOCKED_P "__cpu_simple_lock *lock"
.\"
.Fd "/* obsolete and for ABI compat only -- do not use */"
.Ft void
.Fn __cpu_simple_lock_set "__cpu_simple_Lock *lock"
.Ft void
.Fn __cpu_simple_lock_clear "__cpu_simple_lock *lock"
.Ft int
.Fn __SIMPLELOCK_UNLOCKED_P "__cpu_simple_lock *lock"
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh DESCRIPTION
The
.Nm
functions provide a simple spin-lock facility for limited purposes that
cannot be served by
.Xr mutex 9 ,
such as inside the implementation of
.Xr mutex 9
itself on platforms with limited atomic read/modify/write operations.
.Pp
.Nm
is very limited:
.Bl -bullet
.It
.Nm
provides no debugging or diagnostic support through the
.Dv LOCKDEBUG
option.
.It
.Nm
does not synchronize between code on a CPU and interrupt handlers
running on that CPU \(em you must use it with
.Xr spl 9
for any locks that may be taken in interrupt context; failing to do so
will likely lead to hard-to-debug deadlock.
.It
.Nm
does not block preemption, so a thread holding a lock may be preempted,
potentially requiring other callers to spin for long durations until
the scheduler runs the holder again.
.It
.Nm
does no exponential backoff to reduce memory traffic during
contention.
.El
.Pp
Unless you know what you are doing, you should use
.Xr mutex 9
instead.
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh INITIALIZATION
The macro
.Dv __SIMPLELOCK_UNLOCKED
expands to an initializer for the type
.Vt __cpu_simple_lock_t :
.Dl "__cpu_simple_lock_t lock = __SIMPLELOCK_UNLOCKED;"
.Pp
A
.Vt __cpu_simple_lock_t
object can also be initialized with
.Fn __cpu_simple_lock_init .
.Pp
No actions are needed to destroy a
.Vt __cpu_simple_lock_t
object.
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh FUNCTIONS
.Bl -tag -width abcd
.It Fn __cpu_simple_lock_init lock
Initialize
.Fa lock
for use with the other
.Nm
functions.
.Pp
The caller is responsible for ensuring
.Fn __cpu_simple_lock_init
happens before any use of the other functions.
.Fn __cpu_simple_lock_init
implies no particular memory ordering on its own.
.\""""""""""""""""
.It Fn __cpu_simple_lock lock
Acquire
.Fa lock ,
waiting until it is released if currently held.
.Pp
Any memory operations preceding the previous
.Fn __cpu_simple_unlock
call that released the lock happen before any memory operations after
the next
.Fn __cpu_simple_lock
call that acquires it.
.\""""""""""""""""
.It Fn __cpu_simple_lock_try lock
Try to acquire
.Fa lock ,
without waiting if it is currently held.
Return 1 if successful, 0 if not.
.Pp
Any memory operations preceding the previous
.Fn __cpu_simple_unlock
call that released the lock happen before any memory operations after
the next
.Em successful
.Fn __cpu_simple_lock_try
call that acquires it.
.\""""""""""""""""
.It Fn __cpu_simple_unlock lock
Release
.Fa lock .
.Pp
Any memory operations preceding
.Fn __cpu_simple_unlock
happen before the next call to
.Fn __cpu_simple_lock ,
or the next successful call to
.Fn __cpu_simple_lock_try ,
that acquires
.Fa lock .
.\""""""""""""""""
.It Fn __SIMPLELOCK_LOCKED_P lock
True if
.Fa lock
is currently locked, by anyone.
.Pp
This is normally only used for diagnostic assertions, or for loops
around
.Fn __cpu_simple_lock_try
that also have higher-level functions like blocking interrupts and
performing exponential backoff.
.Pp
No memory ordering is implied.
.El
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh OBSOLETE FUNCTIONS
The following functions abuse the
.Vt __cpu_simple_lock_t
type to store a boolean.
They are used inside the
.Xr pthread 3
library, and were included in the library ABI, so they can't be removed
without breaking the
.Xr pthread 3
ABI.
Do not use these in new code
.Po
except
.Fn __SIMPLELOCK_LOCKED_P
.Pc .
.Bl -tag -width ".Fn __SIMPLELOCK_UNLOCKED_P lock"
.It Fn __cpu_simple_lock_set lock
Set
.Fa lock
to true.
.It Fn __cpu_simple_lock_clear lock
Set
.Fa lock
to false.
.It Fn __SIMPLELOCK_LOCKED_P lock
True iff
.Fa lock
is true.
.It Fn __SIMPLELOCK_UNLOCKED_P lock
True iff
.Fa lock
is false.
.El
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh CODE REFERENCES
The
.Nm
functions are implemented in
.Pa sys/arch/$ARCH/include/lock.h .
.Pp
A machine-independent implementation, using compiler support for
atomic and memory barrier builtins, is available in
.Pa sys/sys/common_lock.h .
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh SEE ALSO
.Xr locking 9 ,
.Xr mutex 9
.\"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
.Sh HISTORY
.Nm
appeared a long time ago.
|