xref: /freebsd/share/man/man9/mi_switch.9 (revision 38f0b757fd84d17d0fc24739a7cda160c4516d81)
1.\"	$NetBSD: ctxsw.9,v 1.2 1996/12/02 00:11:31 tls Exp $
2.\"
3.\" Copyright (c) 1996 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 REGENTS OR CONTRIBUTORS BE
22.\" 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.\" $FreeBSD$
31.\"
32.Dd November 24, 1996
33.Dt MI_SWITCH 9
34.Os
35.Sh NAME
36.Nm mi_switch ,
37.Nm cpu_switch ,
38.Nm cpu_throw
39.Nd switch to another thread context
40.Sh SYNOPSIS
41.In sys/param.h
42.In sys/proc.h
43.Ft void
44.Fn mi_switch "void"
45.Ft void
46.Fn cpu_switch "void"
47.Ft void
48.Fn cpu_throw "void"
49.Sh DESCRIPTION
50The
51.Fn mi_switch
52function implements the machine independent prelude to a thread context
53switch.
54It is called from only a few distinguished places in the kernel
55code as a result of the principle of non-preemptable kernel mode execution.
56The various major uses of
57.Nm
58can be enumerated as follows:
59.Bl -enum -offset indent
60.It
61From within a function such as
62.Xr cv_wait 9 ,
63.Xr mtx_lock 9 ,
64or
65.Xr tsleep 9
66when the current thread
67voluntarily relinquishes the CPU to wait for some resource or lock to become
68available.
69.It
70After handling a trap
71(e.g.\& a system call, device interrupt)
72when the kernel prepares a return to user-mode execution.
73This case is
74typically handled by machine dependent trap-handling code after detection
75of a change in the signal disposition of the current process, or when a
76higher priority thread might be available to run.
77The latter event is
78communicated by the machine independent scheduling routines by calling
79the machine defined
80.Fn need_resched .
81.It
82In the signal handling code
83(see
84.Xr issignal 9 )
85if a signal is delivered that causes a process to stop.
86.It
87When a thread dies in
88.Xr thread_exit 9
89and control of the processor can be passed to the next runnable thread.
90.It
91In
92.Xr thread_suspend_check 9
93where a thread needs to stop execution due to the suspension state of
94the process as a whole.
95.El
96.Pp
97.Fn mi_switch
98records the amount of time the current thread has been running in the
99process structures and checks this value against the CPU time limits
100allocated to the process
101(see
102.Xr getrlimit 2 ) .
103Exceeding the soft limit results in a
104.Dv SIGXCPU
105signal to be posted to the process, while exceeding the hard limit will
106cause a
107.Dv SIGKILL .
108.Pp
109If the thread is still in the
110.Dv TDS_RUNNING
111state,
112.Fn mi_switch
113will put it back onto the run queue, assuming that
114it will want to run again soon.
115If it is in one of the other
116states and KSE threading is enabled, the associated
117.Em KSE
118will be made available to any higher priority threads from the same
119group, to allow them to be scheduled next.
120.Pp
121After these administrative tasks are done,
122.Fn mi_switch
123hands over control to the machine dependent routine
124.Fn cpu_switch ,
125which will perform the actual thread context switch.
126.Pp
127.Fn cpu_switch
128first saves the context of the current thread.
129Next, it calls
130.Fn choosethread
131to determine which thread to run next.
132Finally, it reads in the saved context of the new thread and starts to
133execute the new thread.
134.Pp
135.Fn cpu_throw
136is similar to
137.Fn cpu_switch
138except that it does not save the context of the old thread.
139This function is useful when the kernel does not have an old thread
140context to save, such as when CPUs other than the boot CPU perform their
141first task switch, or when the kernel does not care about the state of the
142old thread, such as in
143.Fn thread_exit
144when the kernel terminates the current thread and switches into a new
145thread.
146.Pp
147To protect the
148.Xr runqueue 9 ,
149all of these functions must be called with the
150.Va sched_lock
151mutex held.
152.Sh SEE ALSO
153.Xr cv_wait 9 ,
154.Xr issignal 9 ,
155.Xr mutex 9 ,
156.Xr runqueue 9 ,
157.Xr tsleep 9 ,
158.Xr wakeup 9
159