xref: /freebsd/share/man/man9/sleep.9 (revision 23f282aa31e9b6fceacd449020e936e98d6f2298)
1.\"
2.\" Copyright (c) 1996 Joerg Wunsch
3.\"
4.\" All rights reserved.
5.\"
6.\" Redistribution and use in source and binary forms, with or without
7.\" modification, are permitted provided that the following conditions
8.\" are met:
9.\" 1. Redistributions of source code must retain the above copyright
10.\"    notice, this list of conditions and the following disclaimer.
11.\" 2. Redistributions in binary form must reproduce the above copyright
12.\"    notice, this list of conditions and the following disclaimer in the
13.\"    documentation and/or other materials provided with the distribution.
14.\"
15.\" THIS SOFTWARE IS PROVIDED BY THE DEVELOPERS ``AS IS'' AND ANY EXPRESS OR
16.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
17.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
18.\" IN NO EVENT SHALL THE DEVELOPERS BE LIABLE FOR ANY DIRECT, INDIRECT,
19.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
20.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
21.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
22.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
23.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
24.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
25.\"
26.\" $FreeBSD$
27.\" "
28.Dd December 17, 1998
29.Os
30.Dt SLEEP 9
31.Sh NAME
32.Nm sleep ,
33.Nm tsleep ,
34.Nm asleep ,
35.Nm await ,
36.Nm wakeup
37.Nd wait for events
38.Sh SYNOPSIS
39.Fd #include <sys/param.h>
40.Fd #include <sys/systm.h>
41.Fd #include <sys/proc.h>
42.Ft int
43.Fn tsleep "void *ident" "int priority" "const char *wmesg" "int timo"
44.Ft int
45.Fn asleep "void *ident" "int priority" "const char *wmesg" "int timo"
46.Ft int
47.Fn await "int priority" "int timo"
48.Ft void
49.Fn wakeup "void *ident"
50.Ft void
51.Fn wakeup_one "void *ident"
52.Sh DESCRIPTION
53The functions
54.Fn tsleep
55and
56.Fn wakeup
57handle event-based process blocking.  If a process must wait for an
58external event, it is put on sleep by
59.Nm tsleep .
60The parameter
61.Ar ident
62is an arbitrary address that uniquely identifies the event on which
63the process is being asleep.  All processes sleeping on a single
64.Ar ident
65are woken up later by
66.Nm wakeup ,
67often called from inside an interrupt routine, to indicate that the
68resource the process was blocking on is available now.
69.Pp
70The parameter
71.Ar wmesg
72is a string describing the sleep condition for tools like
73.Xr ps 1 .
74Due to the limited space of those programs to display arbitrary strings,
75this message should not be longer than 6 characters.
76.Pp
77The
78.Fn wakeup_one
79function is used to make the first process in the queue that is
80sleeping on the parameter
81.Fa ident
82runnable.  This can prevent the system from becoming saturated
83when a large number of processes are sleeping on the same address,
84but only one of them can actually do any useful work when made
85runnable.
86.Pp
87.Nm Tsleep
88is the general sleep call.  Suspends the current process until a wakeup is
89performed on the specified identifier.  The process will then be made
90runnable with the specified
91.Ar priority .
92Sleeps at most
93.Ar timo
94\&/ hz seconds (0 means no timeout).  If
95.Ar pri
96includes the
97.Dv PCATCH
98flag, signals are checked before and after sleeping, else signals are
99not checked.  Returns 0 if awakened,
100.Dv EWOULDBLOCK
101if the timeout expires.  If
102.Dv PCATCH
103is set and a signal needs to be delivered,
104.Dv ERESTART
105is returned if the current system call should be restarted if
106possible, and
107.Dv EINTR
108is returned if the system call should be interrupted by the signal
109.Pq return Dv EINTR .
110.Pp
111.Nm Asleep
112implements the new asynchronous sleep function.  It takes the same arguments
113as
114.Fn tsleep
115and places the process on the appropriate wait queue, but
116.Fn asleep
117leaves the process runnable and returns immediately.  The caller is then
118expected to, at some point in the future, call
119.Fn await
120to actually wait for the previously queued wait condition.
121If
122.Fn asleep
123is called several times, only the most recent call is effective.
124.Fn asleep
125may be called with an
126.Ar ident
127value of NULL
128to remove any previously queued condition.
129.Pp
130.Nm Await
131implements the new asynchronous wait function.  When
132.Fn asleep
133is called on an identifier it associates the process with that
134identifier but does not block.
135.Fn await
136will actually block the process until
137.Fn wakeup
138is called on that identifier any time after the
139.Fn asleep .
140If
141.Fn wakeup
142is called after you
143.Fn asleep
144but before you
145.Fn await
146then the
147.Fn await
148call is effectively a NOP.
149If
150.Fn await
151is called multiple times without an intervening
152.Fn asleep ,
153the
154.Fn await
155is effectively a NOP but will also call
156.Fn mswitch
157for safety.  The
158.Fn await
159function allows you to override the priority and timeout values to be used.
160If the value -1 is specified for an argument, the value is taken from the
161previous
162.Fn asleep
163call.  If -1 is passed for the priority you must be prepared to catch signal
164conditions if the prior call to
165.Fn asleep
166specified it in its priority.  If -1 is passed for the timeout you must be
167prepared to catch a timeout condition if the prior call to
168.Fn asleep
169specified a timeout.  When you use -1, it is usually a good idea to not make
170assumptions as to the arguments used by the prior
171.Fn asleep
172call.
173.Pp
174The
175.Fn asleep
176and
177.Fn await
178functions are mainly used by the kernel to shift the burden of blocking
179away from extremely low level routines and to push it onto their callers.
180This in turn allows more complex interlocking code to
181.Em backout
182of a temporary resource failure
183(such as lack of memory) in order to release major locks prior to actually
184blocking, and to then retry the operation on wakeup.  This key feature is
185expected to be heavily used in SMP situations in order to allow code to make
186better use of spinlocks.  A spinlock, by its very nature, cannot be used
187around code that might block.  It is hoped that these capabilities will
188make it easier to migrate the SMP master locks deeper into the kernel.
189.Pp
190These routines may also be used to avoid nasty spl*() calls to get around
191race conditions with simple conditional test/wait interlocks.  You simply
192call
193.Fn asleep
194prior to your test, then conditionally
195.Fn await
196only if the test fails.  It is usually a good idea to cancel an
197.Fn asleep
198if you wind up never calling the related
199.Fn await ,
200but it is not required.  If you do not want to waste cpu calling
201.Fn asleep
202unnecessarily, you can surround the whole thing with a second test.  The
203race condition is still handled by the inside
204.Fn asleep
205call.
206.Sh RETURN VALUES
207See above.
208.Sh SEE ALSO
209.Xr ps 1 ,
210.Xr malloc 9
211.Sh HISTORY
212The sleep/wakeup process synchronization mechanism is very old.  It
213appeared in a very early version of Unix.
214.Pp
215.Nm Tsleep
216appeared in
217.Bx 4.4 .
218.Pp
219.Nm Asleep/await
220first appeared in
221.Fx 3.0
222and is designed to shift the burden of blocking
223away from extremely low level routines and push it up to their callers.
224.Pp
225.Nm Sleep
226used to be the traditional form.  It doesn't let you specify a timeout nor a
227.Ar wmesg ,
228hence it has been discontinued.
229.Sh AUTHORS
230This man page has been written by
231.ie t J\(:org Wunsch.
232.el Joerg Wunsch.
233.Nm Asleep
234and
235.Nm await
236designed and written by
237.An Matthew Dillon.
238