]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/blob - share/man/man9/mi_switch.9
This commit was generated by cvs2svn to compensate for changes in r76371,
[FreeBSD/FreeBSD.git] / share / man / man9 / mi_switch.9
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 .\" 3. All advertising materials mentioning features or use of this software
18 .\"    must display the following acknowledgement:
19 .\"        This product includes software developed by the NetBSD
20 .\"        Foundation, Inc. and its contributors.
21 .\" 4. Neither the name of The NetBSD Foundation nor the names of its
22 .\"    contributors may be used to endorse or promote products derived
23 .\"    from this software without specific prior written permission.
24 .\"
25 .\" THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
26 .\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
27 .\" TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
28 .\" PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE
29 .\" LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
30 .\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
31 .\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
32 .\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
33 .\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
34 .\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35 .\" POSSIBILITY OF SUCH DAMAGE.
36 .\"
37 .\" $FreeBSD$
38 .\"
39 .Dd November 24, 1996
40 .Dt MI_SWITCH 9
41 .Os FreeBSD
42 .Sh NAME
43 .Nm mi_switch ,
44 .Nm cpu_switch ,
45 .Nm cpu_throw
46 .Nd switch to another process context
47 .Sh SYNOPSIS
48 .Fd #include <sys/param.h>
49 .Fd #include <sys/proc.h>
50 .Ft void
51 .Fn mi_switch "void"
52 .Ft void
53 .Fn cpu_switch "void"
54 .Ft void
55 .Fn cpu_throw "void"
56 .Sh DESCRIPTION
57 The
58 .Fn mi_switch
59 function implements the machine independent prelude to a process context
60 switch.
61 It is called from only a few distinguished places in the kernel
62 code as a result of the principle of non-preemtable kernel mode execution.
63 The three major uses of
64 .Nm
65 can be enumerated as follows:
66 .Bl -enum -offset indent
67 .It
68 from within
69 .Xr sleep 9
70 and
71 .Xr tsleep 9
72 when the current process
73 voluntarily relinquishes the CPU to wait for some resource to become
74 available.
75 .It
76 after handling a trap
77 .Pq e.g. a system call, device interrupt
78 when the kernel prepares a return to user-mode execution.
79 This case is
80 typically handled by machine dependent trap-handling code after detection
81 of a change in the signal disposition of the current process, or when a
82 higher priority process might be available to run.
83 The latter event is
84 communicated by the machine independent scheduling routines by calling
85 the machine defined
86 .Fn need_resched .
87 .It
88 in the signal handling code
89 .Pq see Xr issignal 9
90 if a signal is delivered that causes a process to stop.
91 .El
92 .Pp
93 .Fn mi_switch
94 records the amount of time the current process has been running in the
95 process structure and checks this value against the CPU time limits
96 allocated to the process
97 .Pq see Xr getrlimit 2 .
98 Exceeding the soft limit results in a
99 .Dv SIGXCPU
100 signal to be posted to the process, while exceeding the hard limit will
101 cause a
102 .Dv SIGKILL .
103 After these administrative tasks are done,
104 .Fn mi_switch
105 hands over control to the machine dependent routine
106 .Fn cpu_switch ,
107 which will perform the actual process context switch.
108 .Pp
109 .Fn cpu_switch
110 first saves the context of the current process.
111 Next, it calls
112 .Fn chooseproc
113 to determine which process to run next.
114 Finally, it reads in the saved context of the new process and starts to
115 execute the new process.
116 .Pp
117 .Fn cpu_throw
118 is similar to
119 .Fn cpu_switch
120 except that it does not save the context of the old process.
121 This function is useful when the kernel does not have an old process
122 context to save, such as when CPUs other than the boot CPU perform their
123 first task switch, or when the kernel does not care about the state of the
124 old process, such as in
125 .Fn cpu_exit
126 when the kernel terminates the current process and switches into a new
127 process.
128 .Pp
129 To protect the
130 .Xr runqueue 9 ,
131 all of these functions must be called with the
132 .Va sched_lock
133 mutex held.
134 .Sh SEE ALSO
135 .Xr issignal 9 ,
136 .Xr mutex 9 ,
137 .Xr runqueue 9 ,
138 .Xr tsleep 9 ,
139 .Xr wakeup 9