xref: /netbsd-src/share/man/man9/ltsleep.9 (revision b1c86f5f087524e68db12794ee9c3e3da1ab17a0)
1.\"	$NetBSD: ltsleep.9,v 1.12 2008/04/30 13:10:58 martin Exp $
2.\"
3.\" Copyright (c) 1996, 2002, 2007 The NetBSD Foundation, Inc.
4.\" All rights reserved.
5.\"
6.\" This code is derived from software contributed to The NetBSD Foundation
7.\" by Paul Kranenburg.
8.\"
9.\" Redistribution and use in source and binary forms, with or without
10.\" modification, are permitted provided that the following conditions
11.\" are met:
12.\" 1. Redistributions of source code must retain the above copyright
13.\"    notice, this list of conditions and the following disclaimer.
14.\" 2. Redistributions in binary form must reproduce the above copyright
15.\"    notice, this list of conditions and the following disclaimer in the
16.\"    documentation and/or other materials provided with the distribution.
17.\"
18.\" THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
19.\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
20.\" TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
21.\" PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
22.\" BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23.\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24.\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
25.\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
26.\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
27.\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28.\" POSSIBILITY OF SUCH DAMAGE.
29.\"
30.Dd June 17, 2007
31.Dt LTSLEEP 9
32.Os
33.Sh NAME
34.Nm ltsleep ,
35.Nm tsleep ,
36.Nm wakeup
37.Nd process context sleep and wakeup
38.Sh SYNOPSIS
39.In sys/proc.h
40.Ft int
41.Fn "ltsleep" "wchan_t ident" "pri_t priority" "const char *wmesg" "int timo" "volatile struct simplelock *slock"
42.Ft int
43.Fn "tsleep" "wchan_t ident" "pri_t priority" "const char *wmesg" "int timo"
44.Ft void
45.Fn "wakeup" "wchan_t ident"
46.Sh DESCRIPTION
47.Em The interfaces described in this manual page are obsolete
48.Em and will be removed from a future version of the system.
49.Pp
50.Em Please see the
51.Xr condvar 9 ,
52.Xr mutex 9 ,
53.Em and
54.Xr rwlock 9
55.Em manual pages for information on kernel synchronisation primitives.
56.Pp
57These functions implement voluntary context switching.
58.Fn ltsleep
59and
60.Fn tsleep
61are used throughout the kernel whenever processing in the current context
62can not continue for any of the following reasons:
63.Bl -bullet -offset indent
64.It
65The current process needs to await the results of a pending I/O operation.
66.It
67The current process needs resources
68.Pq e.g., memory
69which are temporarily unavailable.
70.It
71The current process wants access to data-structures which are locked by
72other processes.
73.El
74.Pp
75The function
76.Fn wakeup
77is used to notify sleeping processes of possible changes to the condition
78that caused them to go to sleep.
79Typically, an awakened process will -- after it has acquired a context
80again -- retry the action that blocked its operation to see if the
81.Dq blocking
82condition has cleared.
83.Pp
84The
85.Fn ltsleep
86function takes the following arguments:
87.Bl -tag -width priority
88.It Fa ident
89An identifier of the
90.Dq wait channel
91representing the resource for which the current process needs to wait.
92This typically is the virtual address of some kernel data-structure related
93to the resource for which the process is contending.
94The same identifier must be used in a call to
95.Fn wakeup
96to get the process going again.
97.Fa ident
98should not be
99.Dv NULL .
100.It Fa priority
101The process priority to be used when the process is awakened and put on
102the queue of runnable processes.
103This mechanism is used to optimize
104.Dq throughput
105of processes executing in kernel mode.
106If the flag
107.Dv PCATCH
108is OR'ed into
109.Fa priority
110the process checks for posted signals before and after sleeping.
111If the flag
112.Dv PNORELOCK
113is OR'ed into
114.Fa priority ,
115.Fa slock
116is NOT re-locked after process resume.
117.It Fa wmesg
118A pointer to a character string indicating the reason a process is sleeping.
119The kernel does not use the string, but makes it available
120.Pq through the process structure field Li p_wmesg
121for user level utilities such as
122.Xr ps 1 .
123.It Fa timo
124If non-zero, the process will sleep for at most
125.Li timo/hz
126seconds.
127If this amount of time elapses and no
128.Fn wakeup "ident"
129has occurred, and no signal
130.Pq if Dv PCATCH No was set
131was posted,
132.Fn tsleep
133will return
134.Er EWOULDBLOCK .
135.It Fa slock
136If not NULL, the
137.Fa slock
138interlock is unlocked once the scheduler lock is acquired.
139Unless
140.Dv PNORELOCK
141was set,
142.Fa slock
143is locked again once
144the process is resumed from sleep.
145This provides wakeup-before-sleep condition protection facility.
146.El
147.Pp
148The
149.Fn tsleep
150macro is functionally equivalent to:
151.Bd -literal -offset indent
152ltsleep(ident, priority, wmesg, timo, NULL)
153.Ed
154.Pp
155The
156.Fn wakeup
157function will mark all processes which are currently sleeping on the identifier
158.Fa ident
159as runnable.
160Eventually, each of the processes will resume execution in the kernel
161context, causing a return from
162.Fn tsleep .
163Note that processes returning from sleep should always re-evaluate the
164conditions that blocked them, since a call to
165.Fn wakeup
166merely signals a
167.Em possible
168change to the blocking conditions.
169For example, when two or more processes are waiting for an exclusive-access
170lock
171.Pq see Xr lock 9 ,
172only one of them will succeed in acquiring the lock when it is released.
173All others will have to go back to sleep and wait for the next opportunity.
174.Sh RETURN VALUES
175.Fn ltsleep
176returns 0 if it returns as a result of a
177.Fn wakeup .
178If a
179.Fn ltsleep
180returns as a result of a signal, the return value is
181.Er ERESTART
182if the signal has the
183.Dv SA_RESTART
184property
185.Pq see Xr sigaction 2 ,
186and
187.Er EINTR
188otherwise.
189If
190.Fn ltsleep
191returns because of a timeout it returns
192.Er EWOULDBLOCK .
193.Sh SEE ALSO
194.Xr sigaction 2 ,
195.Xr condvar 9 ,
196.Xr hz 9 ,
197.Xr lock 9 ,
198.Xr mutex 9 ,
199.Xr rwlock 9
200.Sh HISTORY
201The sleep/wakeup process synchronization mechanism is very old.
202It appeared in a very early version of Unix.
203.Fn tsleep
204appeared in
205.Bx 4.4 .
206.Fn ltsleep
207appeared in
208.Nx 1.5 .
209