xref: /netbsd-src/sbin/init/init.8 (revision 466a16a118933bd295a8a104f095714fadf9cf68)
1.\"	$NetBSD: init.8,v 1.55 2008/11/11 06:28:30 wiz Exp $
2.\"
3.\" Copyright (c) 1980, 1991, 1993
4.\"	The Regents of the University of California.  All rights reserved.
5.\"
6.\" This code is derived from software contributed to Berkeley by
7.\" Donn Seeley at Berkeley Software Design, Inc.
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. Neither the name of the University nor the names of its contributors
18.\"    may be used to endorse or promote products derived from this software
19.\"    without specific prior written permission.
20.\"
21.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31.\" SUCH DAMAGE.
32.\"
33.\"     @(#)init.8	8.6 (Berkeley) 5/26/95
34.\"
35.Dd November 10, 2008
36.Dt INIT 8
37.Os
38.Sh NAME
39.Nm init
40.Nd process control initialization
41.Sh SYNOPSIS
42.Nm
43.Sh DESCRIPTION
44The
45.Nm
46program is the last stage of the boot process (after the kernel loads
47and initializes all the devices).
48It normally begins multi-user operation.
49.Pp
50The following table describes the state machine used by
51.Nm :
52.Bl -enum
53.It
54Single user shell.
55.Nm
56may be passed
57.Fl s
58from the boot program to prevent the system from going multi-user and
59to instead execute a single user shell without starting the normal
60daemons.
61If the kernel is in a secure mode,
62.Nm
63will downgrade it to securelevel 0 (insecure mode).
64The system is then quiescent for maintenance work and may
65later be made to go to state 2 (multi-user) by exiting the single-user
66shell (with ^D).
67.It
68Multi-user boot (default operation).
69Executes
70.Pa /etc/rc
71(see
72.Xr rc 8 ) .
73If this was the first state entered (as opposed to entering here after
74state 1), then
75.Pa /etc/rc
76will be invoked with its first argument being
77.Sq autoboot .
78If
79.Pa /etc/rc
80exits with a non-zero (error) exit code, commence single user
81operation by giving the super-user a shell on the console by going
82to state 1 (single user).
83Otherwise, proceed to state 3.
84.Pp
85If value of the
86.Dq init.root
87sysctl node is not equal to
88.Pa /
89at this point, the
90.Pa /etc/rc
91process will be run inside a
92.Xr chroot 2
93indicated by sysctl with the same error handling as above.
94.Pp
95If the administrator has not set the security level to \-1
96to indicate that the kernel should not run multiuser in secure
97mode, and the
98.Pa /etc/rc
99script has not set a higher level of security
100than level 1, then
101.Nm
102will put the kernel into securelevel mode 1.
103See
104.Xr rc.conf 5
105and
106.Xr secmodel_securelevel 9
107for more information.
108.It
109Set up ttys as specified in
110.Xr ttys 5 .
111See below for more information.
112On completion, continue to state 4.
113If we did chroot in state 2, each
114.Xr getty 8
115process will be run in the same
116.Xr chroot 2
117path as in 2 (that is, the value of
118.Dq init.root
119sysctl is not re-read).
120.It
121Multi-user operation.
122Depending upon the signal received, change state appropriately;
123on
124.Dv SIGTERM ,
125go to state 7;
126on
127.Dv SIGHUP ,
128go to state 5;
129on
130.Dv SIGTSTP ,
131go to state 6.
132.It
133Clean-up mode; re-read
134.Xr ttys 5 ,
135killing off the controlling processes on lines that are now
136.Sq off ,
137and starting processes that are newly
138.Sq on .
139On completion, go to state 4.
140.It
141.Sq Boring
142mode; no new sessions.
143Signals as per state 4.
144.It
145Shutdown mode.
146Send
147.Dv SIGHUP
148to all controlling processes, reap the processes for 30 seconds,
149and then go to state 1 (single user); warning if not all the processes died.
150.El
151.Pp
152If the
153.Sq console
154entry in the
155.Xr ttys 5
156file is marked
157.Dq insecure ,
158then
159.Nm
160will require that the superuser password be
161entered before the system will start a single-user shell.
162The password check is skipped if the
163.Sq console
164is marked as
165.Dq secure .
166.Pp
167It should be noted that while
168.Nm
169has the ability to start multi-user operation inside a
170.Xr chroot 2
171environment, the
172.Nm
173process itself will always run in the
174.Dq original root directory .
175This also implies that single-user mode is always started in the original
176root, giving the possibility to create multi-user sessions in different
177root directories over time.
178The
179.Dq init.root
180sysctl node is fabricated by
181.Nm
182at startup and re-created any time it's found to be missing.
183Type of the node is string capable of holding full pathname, and
184is only accessible by the superuser (unless explicitly destroyed
185and re-created with different specification).
186.Pp
187In multi-user operation,
188.Nm
189maintains
190processes for the terminal ports found in the file
191.Xr ttys 5 .
192.Nm
193reads this file, and executes the command found in the second field.
194This command is usually
195.Xr getty 8 ;
196it opens and initializes the tty line and executes the
197.Xr login 1
198program.
199The
200.Xr login 1
201program, when a valid user logs in, executes a shell for that user.
202When this shell dies, either because the user logged out or an
203abnormal termination occurred (a signal), the
204.Nm
205program wakes up, deletes the user from the
206.Xr utmp 5
207and
208.Xr utmpx 5
209files of current users and records the logout in the
210.Xr wtmp 5
211and
212.Xr wtmpx 5
213files.
214The cycle is
215then restarted by
216.Nm
217executing a new
218.Xr getty 8
219for the line.
220.pl +1
221.Pp
222Line status (on, off, secure, getty, or window information)
223may be changed in the
224.Xr ttys 5
225file without a reboot by sending the signal
226.Dv SIGHUP
227to
228.Nm
229with the command
230.Dq Li "kill \-s HUP 1" .
231This is referenced in the table above as state 5.
232On receipt of this signal,
233.Nm
234re-reads the
235.Xr ttys 5
236file.
237When a line is turned off in
238.Xr ttys 5 ,
239.Nm
240will send a
241.Dv SIGHUP
242signal to the controlling process
243for the session associated with the line.
244For any lines that were previously turned off in the
245.Xr ttys 5
246file and are now on,
247.Nm
248executes a new
249.Xr getty 8
250to enable a new login.
251If the getty or window field for a line is changed,
252the change takes effect at the end of the current
253login session (e.g., the next time
254.Nm
255starts a process on the line).
256If a line is commented out or deleted from
257.Xr ttys 5 ,
258.Nm
259will not do anything at all to that line.
260However, it will complain that the relationship between lines
261in the
262.Xr ttys 5
263file and records in the
264.Xr utmp 5
265file is out of sync,
266so this practice is not recommended.
267.Pp
268.Nm
269will terminate multi-user operations and resume single-user mode
270if sent a terminate
271.Pq Dv TERM
272signal, for example,
273.Dq Li "kill \-s TERM 1" .
274If there are processes outstanding that are deadlocked (because of
275hardware or software failure),
276.Nm
277will not wait for them all to die (which might take forever), but
278will time out after 30 seconds and print a warning message.
279.Pp
280.Nm
281will cease creating new
282.Xr getty 8 Ns 's
283and allow the system to slowly die away, if it is sent a terminal stop
284.Pq Dv TSTP
285signal, i.e.
286.Dq Li "kill \-s TSTP 1" .
287A later hangup will resume full
288multi-user operations, or a terminate will start a single user shell.
289This hook is used by
290.Xr reboot 8
291and
292.Xr halt 8 .
293.Pp
294The role of
295.Nm
296is so critical that if it dies, the system will reboot itself
297automatically.
298If, at bootstrap time, the
299.Nm
300process cannot be located, or exits during its initialisation,
301the system will panic with the message
302.Dq panic: init died (signal %d, exit %d) .
303.Pp
304If
305.Pa /dev/console
306does not exist,
307.Nm
308will cd to
309.Pa /dev
310and run
311.Dq Li "MAKEDEV -MM init" .
312.Xr MAKEDEV 8
313will use
314.Xr mount_tmpfs 8
315or
316.Xr mount_mfs 8
317to create a memory file system mounted over
318.Pa /dev
319that contains the standard devices considered necessary to boot the system.
320.Sh FILES
321.Bl -tag -width /var/log/wtmp{,x} -compact
322.It Pa /dev/console
323System console device.
324.It Pa /dev/tty*
325Terminal ports found in
326.Xr ttys 5 .
327.It Pa /var/run/utmp{,x}
328Record of current users on the system.
329.It Pa /var/log/wtmp{,x}
330Record of all logins and logouts.
331.It Pa /etc/ttys
332The terminal initialization information file.
333.It Pa /etc/rc
334System startup commands.
335.El
336.Sh DIAGNOSTICS
337.Bl -diag
338.It "getty repeating too quickly on port %s, sleeping"
339A process being started to service a line is exiting quickly
340each time it is started.
341This is often caused by a ringing or noisy terminal line.
342.Em "Init will sleep for 10 seconds" ,
343.Em "then continue trying to start the process" .
344.Pp
345.It "some processes would not die; ps axl advised."
346A process is hung and could not be killed when the system was
347shutting down.
348This condition is usually caused by a process that is stuck in a
349device driver because of a persistent device error condition.
350.El
351.Sh SEE ALSO
352.Xr config 1 ,
353.Xr kill 1 ,
354.Xr login 1 ,
355.Xr sh 1 ,
356.Xr options 4 ,
357.Xr ttys 5 ,
358.Xr MAKEDEV 8 ,
359.Xr MAKEDEV.local 8 ,
360.Xr getty 8 ,
361.Xr halt 8 ,
362.Xr mount_mfs 8 ,
363.Xr mount_tmpfs 8 ,
364.Xr rc 8 ,
365.Xr reboot 8 ,
366.Xr rescue 8 ,
367.Xr shutdown 8 ,
368.Xr sysctl 8 ,
369.Xr secmodel_bsd44 9 ,
370.Xr secmodel_securelevel 9
371.Sh HISTORY
372A
373.Nm
374command appeared in
375.At v6 .
376