2 The FreeBSD Documentation Project
4 $FreeBSD: doc/en_US.ISO8859-1/books/handbook/config/chapter.sgml,v 1.154 2004/06/08 11:41:49 den Exp $
5 $DragonFly: doc/en/books/handbook/config/chapter.sgml,v 1.3 2004/08/25 20:42:15 justin Exp $
8 <chapter id="config-tuning">
12 <firstname>Chern</firstname>
13 <surname>Lee</surname>
14 <contrib>Written by </contrib>
19 <firstname>Mike</firstname>
20 <surname>Smith</surname>
21 <contrib>Based on a tutorial written by </contrib>
26 <firstname>Matt</firstname>
27 <surname>Dillon</surname>
28 <contrib>Also based on tuning(7) written by </contrib>
33 <title>Configuration and Tuning</title>
35 <sect1 id="config-synopsis">
36 <title>Synopsis</title>
38 <indexterm><primary>system configuration</primary></indexterm>
39 <indexterm><primary>system optimization</primary></indexterm>
41 <para>One of the important aspects of &os; is system configuration.
42 Correct system configuration will help prevent headaches during future upgrades.
43 This chapter will explain much of the &os; configuration process,
44 including some of the parameters which
45 can be set to tune a &os; system.
48 <para>After reading this chapter, you will know:</para>
52 <para>How to efficiently work with
53 file systems and swap partitions.</para>
56 <para>The basics of <filename>rc.conf</filename> configuration and
57 <filename>/usr/local/etc/rc.d</filename> startup systems.</para>
60 <para>How to configure and test a network card.</para>
63 <para>How to configure virtual hosts on your network devices.</para>
66 <para>How to use the various configuration files in
67 <filename>/etc</filename>.</para>
70 <para>How to tune &os; using <command>sysctl</command>
74 <para>How to tune disk performance and modify kernel
79 <para>Before reading this chapter, you should:</para>
83 <para>Understand &unix; and &os; basics (<xref
84 linkend="basics">).</para>
87 <para>Be familiar with keeping &os; sources up to date
88 (<xref linkend="updating-setup">), and
89 the basics of kernel configuration/compilation
90 (<xref linkend="kernelconfig">).</para>
95 <sect1 id="configtuning-initial">
96 <title>Initial Configuration</title>
99 <title>Partition Layout</title>
101 <indexterm><primary>partition layout</primary></indexterm>
103 <primary><filename>/etc</filename></primary>
106 <primary><filename>/var</filename></primary>
109 <primary><filename>/usr</filename></primary>
113 <title>Base Partitions</title>
115 <para>When laying out file systems with &man.disklabel.8;
116 or &man.sysinstall.8;, remember that hard
117 drives transfer data faster from the outer
119 Thus smaller and heavier-accessed file systems
120 should be closer to the outside of the drive, while
121 larger partitions like <filename>/usr</filename> should be placed
122 toward the inner. It is a good idea to create
123 partitions in a similar order to: root, swap,
124 <filename>/var</filename>, <filename>/usr</filename>.</para>
126 <para>The size of <filename>/var</filename>
127 reflects the intended machine usage.
128 <filename>/var</filename> is used to hold
129 mailboxes, log files, and printer spools. Mailboxes and log
130 files can grow to unexpected sizes depending
131 on how many users exist and how long log
132 files are kept. Most users would never require a gigabyte,
133 but remember that <filename>/var/tmp</filename>
134 must be large enough to contain packages.
137 <para>The <filename>/usr</filename> partition holds much
138 of the files required to support the system, the &man.ports.7;
139 collection (recommended) and the source code (optional). Both
140 of which are optional at install time.
141 At least 2 gigabytes would be recommended for this partition.</para>
143 <para>When selecting partition sizes, keep the space
144 requirements in mind. Running out of space in
145 one partition while barely using another can be a
148 <note><para>Some users have found that &man.sysinstall.8;'s
149 <literal>Auto-defaults</literal> partition sizer will
150 sometimes select smaller than adequate <filename>/var</filename>
151 and <filename>/</filename> partitions. Partition wisely and
152 generously.</para></note>
156 <sect3 id="swap-design">
157 <title>Swap Partition</title>
159 <indexterm><primary>swap sizing</primary></indexterm>
160 <indexterm><primary>swap partition</primary></indexterm>
162 <para>As a rule of thumb, the swap partition should be
163 about double the size of system memory (RAM). For example,
164 if the machine has 128 megabytes of memory,
165 the swap file should be 256 megabytes. Systems with
166 less memory may perform better with more swap.
167 Less than 256 megabytes of swap is not recommended and
168 memory expansion should be considered.
169 The kernel's VM paging algorithms are tuned to
170 perform best when the swap partition is at least two times the
171 size of main memory. Configuring too little swap can lead to
172 inefficiencies in the VM page scanning code and might create
173 issues later if more memory is added.</para>
175 <para>On larger systems with multiple SCSI disks (or
176 multiple IDE disks operating on different controllers), it is
177 recommend that a swap is configured on each drive (up
178 to four drives). The swap partitions should be
179 approximately the same size. The kernel can handle arbitrary
180 sizes but internal data structures scale to 4 times the
181 largest swap partition. Keeping the swap partitions near the
182 same size will allow the kernel to optimally stripe swap space
184 Large swap sizes are fine, even if swap is not
185 used much. It might be easier to recover
186 from a runaway program before being forced to reboot.</para>
190 <title>Why Partition?</title>
192 <para>Several users think a single large partition will be fine,
193 but there are several reasons why this is a bad idea.
194 First, each partition has different operational
195 characteristics and separating them allows the file system to
196 tune accordingly. For example, the root
197 and <filename>/usr</filename> partitions are read-mostly, without
198 much writing. While a lot of reading and writing could
199 occur in <filename>/var</filename> and
200 <filename>/var/tmp</filename>.</para>
202 <para>By properly partitioning a system, fragmentation
203 introduced in the smaller write heavy partitions
204 will not bleed over into the mostly-read partitions.
205 Keeping the write-loaded partitions closer to
208 increase I/O performance in the partitions where it occurs
209 the most. Now while I/O
210 performance in the larger partitions may be needed,
211 shifting them more toward the edge of the disk will not
212 lead to a significant performance improvement over moving
213 <filename>/var</filename> to the edge.
214 Finally, there are safety concerns. A smaller, neater root
215 partition which is mostly read-only has a greater
216 chance of surviving a bad crash.</para>
222 <sect1 id="configtuning-core-configuration">
223 <title>Core Configuration</title>
226 <primary>rc files</primary>
227 <secondary><filename>rc.conf</filename></secondary>
230 <para>The principal location for system configuration information
231 is within <filename>/etc/rc.conf</filename>. This file
232 contains a wide range of configuration information, principally
233 used at system startup to configure the system. Its name
234 directly implies this; it is configuration information for the
235 <filename>rc*</filename> files.</para>
237 <para>An administrator should make entries in the
238 <filename>rc.conf</filename> file to
239 override the default settings from
240 <filename>/etc/defaults/rc.conf</filename>. The defaults file
241 should not be copied verbatim to <filename>/etc</filename> - it
242 contains default values, not examples. All system-specific
243 changes should be made in the <filename>rc.conf</filename>
246 <para>A number of strategies may be applied in clustered
247 applications to separate site-wide configuration from
248 system-specific configuration in order to keep administration
249 overhead down. The recommended approach is to place site-wide
250 configuration into another file,
251 such as <filename>/etc/rc.conf.site</filename>, and then include
252 this file into <filename>/etc/rc.conf</filename>, which will
253 contain only system-specific information.</para>
255 <para>As <filename>rc.conf</filename> is read by &man.sh.1; it is
256 trivial to achieve this. For example:</para>
259 <listitem><para>rc.conf:</para>
260 <programlisting> . rc.conf.site
261 hostname="node15.example.com"
262 network_interfaces="fxp0 lo0"
263 ifconfig_fxp0="inet 10.1.1.1"</programlisting></listitem>
264 <listitem><para>rc.conf.site:</para>
265 <programlisting> defaultrouter="10.1.1.254"
267 blanktime="100"</programlisting></listitem>
270 <para>The <filename>rc.conf.site</filename> file can then be
271 distributed to every system using <command>rsync</command> or a
272 similar program, while the <filename>rc.conf</filename> file
273 remains unique.</para>
275 <para>Upgrading the system using &man.sysinstall.8;
276 or <command>make world</command> will not overwrite the
277 <filename>rc.conf</filename>
278 file, so system configuration information will not be lost.</para>
282 <sect1 id="configtuning-appconfig">
283 <title>Application Configuration</title>
285 <para>Typically, installed applications have their own
286 configuration files, with their own syntax, etc. It is
287 important that these files be kept separate from the base
288 system, so that they may be easily located and managed by the
289 package management tools.</para>
291 <indexterm><primary>/usr/local/etc</primary></indexterm>
293 <para>Typically, these files are installed in
294 <filename>/usr/local/etc</filename>. In the case where an
295 application has a large number of configuration files, a
296 subdirectory will be created to hold them.</para>
298 <para>Normally, when a port or package is installed, sample
299 configuration files are also installed. These are usually
300 identified with a <filename>.default</filename> suffix. If there
302 configuration files for the application, they will be created by
303 copying the <filename>.default</filename> files.</para>
305 <para>For example, consider the contents of the directory
306 <filename>/usr/local/etc/apache</filename>:</para>
308 <literallayout class="monospaced">-rw-r--r-- 1 root wheel 2184 May 20 1998 access.conf
309 -rw-r--r-- 1 root wheel 2184 May 20 1998 access.conf.default
310 -rw-r--r-- 1 root wheel 9555 May 20 1998 httpd.conf
311 -rw-r--r-- 1 root wheel 9555 May 20 1998 httpd.conf.default
312 -rw-r--r-- 1 root wheel 12205 May 20 1998 magic
313 -rw-r--r-- 1 root wheel 12205 May 20 1998 magic.default
314 -rw-r--r-- 1 root wheel 2700 May 20 1998 mime.types
315 -rw-r--r-- 1 root wheel 2700 May 20 1998 mime.types.default
316 -rw-r--r-- 1 root wheel 7980 May 20 1998 srm.conf
317 -rw-r--r-- 1 root wheel 7933 May 20 1998 srm.conf.default</literallayout>
319 <para>The file sizes show that only the <filename>srm.conf</filename>
320 file has been changed. A later update of the <application>Apache</application> port would not
321 overwrite this changed file.</para>
325 <sect1 id="configtuning-starting-services">
326 <title>Starting Services</title>
328 <indexterm><primary>services</primary></indexterm>
330 <para>It is common for a system to host a number of services.
331 These may be started in several different fashions, each having
332 different advantages.</para>
334 <indexterm><primary>/usr/local/etc/rc.d</primary></indexterm>
336 <para>Software installed from a port or the packages collection
337 will often place a script in
338 <filename>/usr/local/etc/rc.d</filename> which is invoked at
339 system startup with a <option>start</option> argument, and at
340 system shutdown with a <option>stop</option> argument.
341 This is the recommended way for
342 starting system-wide services that are to be run as
343 <username>root</username>, or that
344 expect to be started as <username>root</username>.
345 These scripts are registered as
346 part of the installation of the package, and will be removed
347 when the package is removed.</para>
349 <para>A generic startup script in
350 <filename>/usr/local/etc/rc.d</filename> looks like:</para>
352 <programlisting>#!/bin/sh
357 /usr/local/bin/foobar
360 kill -9 `cat /var/run/foobar.pid`
363 echo "Usage: `basename $0` {start|stop}" >&2
371 <para>The startup scripts of &os; will look in
372 <filename>/usr/local/etc/rc.d</filename> for scripts that have an
373 <literal>.sh</literal> extension and are executable by
374 <username>root</username>. Those scripts that are found are called with
375 an option <option>start</option> at startup, and <option>stop</option>
376 at shutdown to allow them to carry out their purpose. So if you wanted
377 the above sample script to be picked up and run at the proper time during
378 system startup, you should save it to a file called
379 <filename>FooBar.sh</filename> in
380 <filename>/usr/local/etc/rc.d</filename> and make sure it is
381 executable. You can make a shell script executable with &man.chmod.1;
382 as shown below:</para>
384 <screen>&prompt.root; <userinput>chmod 755 <replaceable>FooBar.sh</replaceable></userinput></screen>
386 <para>Some services expect to be invoked by &man.inetd.8; when a
387 connection is received on a suitable port. This is common for
388 mail reader servers (POP and IMAP, etc.). These services are
389 enabled by editing the file <filename>/etc/inetd.conf</filename>.
390 See &man.inetd.8; for details on editing this file.</para>
392 <para>Some additional system services may not be covered by the
393 toggles in <filename>/etc/rc.conf</filename>. These are
394 traditionally enabled by placing the command(s) to invoke them
395 in <filename>/etc/rc.local</filename>. As of &os; 3.1 there
396 is no default <filename>/etc/rc.local</filename>; if it is
397 created by the administrator it will however be honored in the
398 normal fashion. Note that <filename>rc.local</filename> is
399 generally regarded as the location of last resort; if there is a
400 better place to start a service, do it there.</para>
402 <note><para>Do <emphasis>not</emphasis> place any commands in
403 <filename>/etc/rc.conf</filename>. To start daemons, or
404 run any commands at boot time, place a script in
405 <filename>/usr/local/etc/rc.d</filename> instead.</para>
408 <para>It is also possible to use the &man.cron.8; daemon to start
409 system services. This approach has a number of advantages, not
410 least being that because &man.cron.8; runs these processes as the
411 owner of the <command>crontab</command>, services may be started
412 and maintained by non-<username>root</username> users.</para>
414 <para>This takes advantage of a feature of &man.cron.8;: the
415 time specification may be replaced by <literal>@reboot</literal>,
417 cause the job to be run when &man.cron.8; is started shortly after
421 <sect1 id="configtuning-cron">
425 <firstname>Tom</firstname>
426 <surname>Rhodes</surname>
427 <contrib>Contributed by </contrib>
432 <title>Configuring the <command>cron</command> Utility</title>
434 <indexterm><primary>cron</primary>
435 <secondary>configuration</secondary></indexterm>
437 <para>One of the most useful utilities in &os; is &man.cron.8;. The
438 <command>cron</command> utility runs in the background and constantly
439 checks the <filename>/etc/crontab</filename> file. The <command>cron</command>
440 utility also checks the <filename>/var/cron/tabs</filename> directory, in
441 search of new <filename>crontab</filename> files. These
442 <filename>crontab</filename> files store information about specific
443 functions which <command>cron</command> is supposed to perform at
444 certain times.</para>
446 <para>The <command>cron</command> utility uses two different
447 types of configuration files, the system crontab and user crontabs. The
448 only difference between these two formats is the sixth field. In the
449 system crontab, the sixth field is the name of a user for the command
450 to run as. This gives the system crontab the ability to run commands
451 as any user. In a user crontab, the sixth field is the command to run,
452 and all commands run as the user who created the crontab; this is an
453 important security feature.</para>
456 <para>User crontabs allow individual users to schedule tasks without the
457 need for root privileges. Commands in a user's crontab run with the
458 permissions of the user who owns the crontab.</para>
460 <para>The <username>root</username> user can have a user crontab just like
461 any other user. This one is different from
462 <filename>/etc/crontab</filename> (the system crontab). Because of the
463 system crontab, there's usually no need to create a user crontab
464 for <username>root</username>.</para>
467 <para>Let us take a look at the <filename>/etc/crontab</filename> file
468 (the system crontab):</para>
471 <programlisting># /etc/crontab - root's crontab for &os;
473 # $&os;: src/etc/crontab,v 1.32 2002/11/22 16:13:39 tom Exp $
474 # <co id="co-comments">
477 PATH=/etc:/bin:/sbin:/usr/bin:/usr/sbin <co id="co-env">
481 #minute hour mday month wday who command <co id="co-field-descr">
484 */5 * * * * root /usr/libexec/atrun <co id="co-main">
488 <callout arearefs="co-comments">
489 <para>Like most &os; configuration files, the <literal>#</literal>
490 character represents a comment. A comment can be placed in
491 the file as a reminder of what and why a desired action is performed.
492 Comments cannot be on the same line as a command or else they will
493 be interpreted as part of the command; they must be on a new line.
494 Blank lines are ignored.</para>
497 <callout arearefs="co-env">
498 <para>First, the environment must be defined. The equals
499 (<literal>=</literal>) character is used to define any environment
500 settings, as with this example where it is used for the <envar>SHELL</envar>,
501 <envar>PATH</envar>, and <envar>HOME</envar> options. If the shell line is
502 omitted, <command>cron</command> will use the default, which is
503 <command>sh</command>. If the <envar>PATH</envar> variable is
504 omitted, no default will be used and file locations will need to
505 be absolute. If <envar>HOME</envar> is omitted, <command>cron</command>
506 will use the invoking users home directory.</para>
509 <callout arearefs="co-field-descr">
510 <para>This line defines a total of seven fields. Listed here are the
511 values <literal>minute</literal>, <literal>hour</literal>,
512 <literal>mday</literal>, <literal>month</literal>, <literal>wday</literal>,
513 <literal>who</literal>, and <literal>command</literal>. These
514 are almost all self explanatory. <literal>minute</literal> is the time in minutes the
515 command will be run. <literal>hour</literal> is similar to the <literal>minute</literal> option, just in
516 hours. <literal>mday</literal> stands for day of the month. <literal>month</literal> is similar to <literal>hour</literal>
517 and <literal>minute</literal>, as it designates the month. The <literal>wday</literal> option stands for
518 day of the week. All these fields must be numeric values, and follow
519 the twenty-four hour clock. The <literal>who</literal> field is special,
520 and only exists in the <filename>/etc/crontab</filename> file.
521 This field specifies which user the command should be run as.
522 When a user installs his or her <filename>crontab</filename> file, they
523 will not have this option. Finally, the <literal>command</literal> option is listed.
524 This is the last field, so naturally it should designate the command
525 to be executed.</para>
528 <callout arearefs="co-main">
529 <para>This last line will define the values discussed above. Notice here
530 we have a <literal>*/5</literal> listing, followed by several more
531 <literal>*</literal> characters. These <literal>*</literal> characters
532 mean <quote>first-last</quote>, and can be interpreted as
533 <emphasis>every</emphasis> time. So, judging by this line,
534 it is apparent that the <command>atrun</command> command is to be invoked by
535 <username>root</username> every five minutes regardless of what
536 day or month it is. For more information on the <command>atrun</command> command,
537 see the &man.atrun.8; manual page.</para>
539 <para>Commands can have any number of flags passed to them; however,
540 commands which extend to multiple lines need to be broken with the backslash
541 <quote>\</quote> continuation character.</para>
545 <para>This is the basic set up for every
546 <filename>crontab</filename> file, although there is one thing
547 different about this one. Field number six, where we specified
548 the username, only exists in the system
549 <filename>/etc/crontab</filename> file. This field should be
550 omitted for individual user <filename>crontab</filename>
554 <sect2 id="configtuning-installcrontab">
555 <title>Installing a Crontab</title>
558 <para>You must not use the procedure described here to
559 edit/install the system crontab. Simply use your favorite
560 editor: the <command>cron</command> utility will notice that the file
561 has changed and immediately begin using the updated version.
563 <ulink url="http://www.freebsd.org/doc/en_US.ISO8859-1/books/faq/admin.html#ROOT-NOT-FOUND-CRON-ERRORS">
564 this FAQ entry </ulink> for more information.</para>
567 <para>To install a freshly written user
568 <filename>crontab</filename>, first use your favorite editor to create
569 a file in the proper format, and then use the
570 <command>crontab</command> utility. The most common usage
573 <screen>&prompt.user; <userinput>crontab crontab-file</userinput></screen>
575 <para>In this example, <filename>crontab-file</filename> is the filename
576 of a <filename>crontab</filename> that was previously created.</para>
578 <para>There is also an option to list installed
579 <filename>crontab</filename> files: just pass the
580 <option>-l</option> option to <command>crontab</command> and look
581 over the output.</para>
583 <para>For users who wish to begin their own crontab file from scratch,
584 without the use of a template, the <command>crontab -e</command>
585 option is available. This will invoke the selected editor
586 with an empty file. When the file is saved, it will be
587 automatically installed by the <command>crontab</command> command.
590 <para>If you later want to remove your user <filename>crontab</filename>
591 completely, use <command>crontab</command> with the <option>-r</option>
598 <sect1 id="configtuning-rcNG">
602 <firstname>Tom</firstname>
603 <surname>Rhodes</surname>
604 <contrib>Contributed by </contrib>
610 <title>Using rc under &os;</title>
612 <indexterm><primary>rcNG</primary></indexterm>
614 <para>&os; uses the NetBSD
615 <filename>rc.d</filename> system for system initialization.
616 Users should notice the files listed in the
617 <filename>/etc/rc.d</filename> directory. Many of these files
618 are for basic services which can be controlled with the
619 <option>start</option>, <option>stop</option>,
620 and <option>restart</option> options.
621 For instance, &man.sshd.8; can be restarted with the following
624 <screen>&prompt.root; <userinput>/etc/rc.d/sshd restart</userinput></screen>
626 <para>This procedure is similar for other services. Of course,
627 services are usually started automatically as specified in
628 &man.rc.conf.5;. For example, enabling the Network Address
629 Translation daemon at startup is as simple as adding the
630 following line to <filename>/etc/rc.conf</filename>:</para>
632 <programlisting>natd_enable="YES"</programlisting>
634 <para>If a <option>natd_enable="NO"</option> line is already
635 present, then simply change the <option>NO</option> to
636 <option>YES</option>. The rc scripts will automatically load
637 any other dependent services during the next reboot, as
638 described below.</para>
640 <para>Since the <filename>rc.d</filename> system is primarily
641 intended to start/stop services at system startup/shutdown time,
642 the standard <option>start</option>,
643 <option>stop</option> and <option>restart</option> options will only
644 perform their action if the appropriate
645 <filename>/etc/rc.conf</filename> variables are set. For
646 instance the above <command>sshd restart</command> command will
647 only work if <varname>sshd_enable</varname> is set to
648 <option>YES</option> in <filename>/etc/rc.conf</filename>. To
649 <option>start</option>, <option>stop</option> or
650 <option>restart</option> a service regardless of the settings in
651 <filename>/etc/rc.conf</filename>, the commands should be
652 prefixed with <quote>force</quote>. For instance to restart
653 <command>sshd</command> regardless of the current
654 <filename>/etc/rc.conf</filename> setting, execute the following
657 <screen>&prompt.root; <userinput>/etc/rc.d/sshd forcerestart</userinput></screen>
659 <para>It is easy to check if a service is enabled in
660 <filename>/etc/rc.conf</filename> by running the appropriate
661 <filename>rc.d</filename> script with the option
662 <option>rcvar</option>. Thus, an administrator can check that
663 <command>sshd</command> is in fact enabled in
664 <filename>/etc/rc.conf</filename> by running:</para>
666 <screen>&prompt.root; <userinput>/etc/rc.d/sshd rcvar</userinput>
668 $sshd_enable=YES</screen>
671 <para>The second line (<literal># sshd</literal>) is the output
672 from the <command>sshd</command> command, not a <username>root</username>
676 <para>To determine if a service is running, a
677 <option>status</option> option is available. For instance to
678 verify that <command>sshd</command> is actually started:</para>
680 <screen>&prompt.root; <userinput>/etc/rc.d/sshd status</userinput>
681 sshd is running as pid 433.</screen>
683 <para>It is also possible to <option>reload</option> a service.
684 This will attempt to send a signal to an individual service, forcing the
685 service to reload its configuration files. In most cases this
686 means sending the service a <literal>SIGHUP</literal>
689 <para>The <application>rcNG</application> structure is used both
690 for network services and system initialization. Some services are run
691 only at boot; and the RCNG system is what triggers them.
693 <para>Many system services depend on other services to function
694 properly. For example, NIS and other RPC-based services may
695 fail to start until after the <command>rpcbind</command>
696 (portmapper) service has started. To resolve this issue,
697 information about dependencies and other meta-data is included
698 in the comments at the top of each startup script. The
699 &man.rcorder.8; program is then used to parse these comments
700 during system initialization to determine the order in which
701 system services should be invoked to satisfy the dependencies.
702 The following words may be included at the top of each startup
707 <para><literal>PROVIDE</literal>: Specifies the services this file provides.</para>
711 <para><literal>REQUIRE</literal>: Lists services which are required for this
712 service. This file will run <emphasis>after</emphasis>
713 the specified services.</para>
717 <para><literal>BEFORE</literal>: Lists services which depend on this service.
718 This file will run <emphasis>before</emphasis>
719 the specified services.</para>
723 <para>KEYWORD: &os; or NetBSD. This is used for *BSD dependent features.</para>
727 <para>By using this method, an administrator can easily control system
728 services without the hassle of <quote>runlevels</quote> like
729 some other &unix; operating systems.</para>
731 <para>Additional information about the &os;
732 <filename>rc.d</filename> system can be found in the &man.rc.8;
733 and &man.rc.subr.8; manual pages.</para>
736 <sect1 id="config-network-setup">
740 <firstname>Marc</firstname>
741 <surname>Fonvieille</surname>
742 <contrib>Contributed by </contrib>
743 <!-- 6 October 2002 -->
748 <title>Setting Up Network Interface Cards</title>
750 <indexterm><primary>network card configuration</primary></indexterm>
752 <para>Nowadays we can not think about a computer without thinking
753 about a network connection. Adding and configuring a network
754 card is a common task for any &os; administrator.</para>
757 <title>Locating the Correct Driver</title>
760 <primary>network card configuration</primary>
761 <secondary>locating the driver</secondary>
764 <para>Before you begin, you should know the model of the card
765 you have, the chip it uses, and whether it is a PCI or ISA card.
766 &os; supports a wide variety of both PCI and ISA cards.
767 Check the Hardware Compatibility List for your release to see
768 if your card is supported.</para>
770 <para>Once you are sure your card is supported, you need
771 to determine the proper driver for the card. The file
772 <filename>/usr/src/sys/i386/conf/LINT</filename> will give you
773 the list of network interfaces drivers with some information
774 about the supported chipsets/cards. If you have doubts about
775 which driver is the correct one, read the manual page of the
776 driver. The manual page will give you more information about
777 the supported hardware and even the possible problems that
780 <para>If you own a common card, most of the time you will not
781 have to look very hard for a driver. Drivers for common
782 network cards are present in the <filename>GENERIC</filename>
783 kernel, so your card should show up during boot, like so:</para>
785 <screen>dc0: <82c169 PNIC 10/100BaseTX> port 0xa000-0xa0ff mem 0xd3800000-0xd38
786 000ff irq 15 at device 11.0 on pci0
787 dc0: Ethernet address: 00:a0:cc:da:da:da
788 miibus0: <MII bus> on dc0
789 ukphy0: <Generic IEEE 802.3u media interface> on miibus0
790 ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
791 dc1: <82c169 PNIC 10/100BaseTX> port 0x9800-0x98ff mem 0xd3000000-0xd30
792 000ff irq 11 at device 12.0 on pci0
793 dc1: Ethernet address: 00:a0:cc:da:da:db
794 miibus1: <MII bus> on dc1
795 ukphy1: <Generic IEEE 802.3u media interface> on miibus1
796 ukphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto</screen>
798 <para>In this example, we see that two cards using the &man.dc.4;
799 driver are present on the system.</para>
801 <para>To use your network card, you will need to load the proper
802 driver. This may be accomplished in one of two ways. The
803 easiest way is to simply load a kernel module for your network
804 card with &man.kldload.8;. A module is not available for all
805 network card drivers (ISA cards and cards using the &man.ed.4;
806 driver, for example). Alternatively, you may statically compile
807 the support for your card into your kernel. Check
808 <filename>/usr/src/sys/i386/conf/LINT</filename> and the
809 manual page of the driver to know what to add in your kernel
810 configuration file. For more information about recompiling your
811 kernel, please see <xref linkend="kernelconfig">. If your card
812 was detected at boot by your kernel (<filename>GENERIC</filename>)
813 you do not have to build a new kernel.</para>
817 <title>Configuring the Network Card</title>
820 <primary>Network card configuration</primary>
821 <secondary>configuration</secondary>
824 <para>Once the right driver is loaded for the network card, the
825 card needs to be configured. As with many other things, the
826 network card may have been configured at installation time by
827 <application>sysinstall</application>.</para>
829 <para>To display the configuration for the network interfaces on
830 your system, enter the following command:</para>
832 <screen>&prompt.user; <userinput>ifconfig</userinput>
833 dc0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
834 inet 192.168.1.3 netmask 0xffffff00 broadcast 192.168.1.255
835 ether 00:a0:cc:da:da:da
836 media: Ethernet autoselect (100baseTX <full-duplex>)
838 dc1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
839 inet 10.0.0.1 netmask 0xffffff00 broadcast 10.0.0.255
840 ether 00:a0:cc:da:da:db
841 media: Ethernet 10baseT/UTP
843 lp0: flags=8810<POINTOPOINT,SIMPLEX,MULTICAST> mtu 1500
844 lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 16384
845 inet 127.0.0.1 netmask 0xff000000
846 tun0: flags=8010<POINTOPOINT,MULTICAST> mtu 1500</screen>
849 <para>Note that entries concerning IPv6
850 (<literal>inet6</literal> etc.) were omitted in this
854 <para>In this example, the following devices were
859 <para><devicename>dc0</devicename>: The first Ethernet
864 <para><devicename>dc1</devicename>: The second Ethernet
869 <para><devicename>lp0</devicename>: The parallel port
874 <para><devicename>lo0</devicename>: The loopback device</para>
878 <para><devicename>tun0</devicename>: The tunnel device used by
879 <application>ppp</application></para>
883 <para>&os; uses the driver name followed by the order in
884 which one the card is detected at the kernel boot to name the
885 network card, starting the count at zero. For example,
886 <devicename>sis2</devicename> would be the third network card
887 on the system using the &man.sis.4; driver.</para>
889 <para>In this example, the <devicename>dc0</devicename> device is
890 up and running. The key indicators are:</para>
894 <para><literal>UP</literal> means that the card is configured
899 <para>The card has an Internet (<literal>inet</literal>)
900 address (in this case
901 <hostid role="ipaddr">192.168.1.3</hostid>).</para>
905 <para>It has a valid subnet mask (<literal>netmask</literal>;
906 <hostid role="netmask">0xffffff00</hostid> is the same as
907 <hostid role="netmask">255.255.255.0</hostid>).</para>
911 <para>It has a valid broadcast address (in this case,
912 <hostid role="ipaddr">192.168.1.255</hostid>).</para>
916 <para>The MAC address of the card (<literal>ether</literal>)
917 is <hostid role="mac">00:a0:cc:da:da:da</hostid></para>
921 <para>The physical media selection is on autoselection mode
922 (<literal>media: Ethernet autoselect (100baseTX
923 <full-duplex>)</literal>). We see that
924 <devicename>dc1</devicename> was configured to run with
925 <literal>10baseT/UTP</literal> media. For more
926 information on available media types for a driver, please
927 refer to its manual page.</para>
931 <para>The status of the link (<literal>status</literal>)
932 is <literal>active</literal>, i.e. the carrier is detected.
933 For <devicename>dc1</devicename>, we see
934 <literal>status: no carrier</literal>. This is normal when
935 an ethernet cable is not plugged into the card.</para>
939 <para>If the &man.ifconfig.8; output had shown something similar
942 <screen>dc0: flags=8843<BROADCAST,SIMPLEX,MULTICAST> mtu 1500
943 ether 00:a0:cc:da:da:da</screen>
945 <para>it would indicate the card has not been configured.</para>
947 <para>To configure your card, you need <username>root</username>
948 privileges. The network card configuration can be done from the
949 command line with &man.ifconfig.8; as root.</para>
952 &prompt.root; <userinput>ifconfig dc0 inet 192.168.1.3 netmask 255.255.255.0</userinput>
956 <para>Manually configuring the care has the disadvantage that you
957 would have to do it after each reboot of the system. The file
958 <filename>/etc/rc.conf</filename> is where to add the network
959 card's configuration.</para>
961 <para>Open <filename>/etc/rc.conf</filename> in your favorite
962 editor. You need to add a line for each network card present on
963 the system, for example in our case, we added these lines:</para>
965 <programlisting>ifconfig_dc0="inet 192.168.1.3 netmask 255.255.255.0"
966 ifconfig_dc1="inet 10.0.0.1 netmask 255.255.255.0 media 10baseT/UTP"</programlisting>
968 <para>You have to replace <devicename>dc0</devicename>,
969 <devicename>dc1</devicename>, and so on, with
970 the correct device for your cards, and the addresses with the
971 proper ones. You should read the card driver and
972 &man.ifconfig.8; manual pages for more details about the allowed
973 options and also &man.rc.conf.5; manual page for more
974 information on the syntax of
975 <filename>/etc/rc.conf</filename>.</para>
977 <para>If you configured the network during installation, some
978 lines about the network card(s) may be already present. Double
979 check <filename>/etc/rc.conf</filename> before adding any
982 <para>You will also have to edit the file
983 <filename>/etc/hosts</filename> to add the names and the IP
984 addresses of various machines of the LAN, if they are not already
985 there. For more information please refer to &man.hosts.5;
986 and to <filename>/usr/share/examples/etc/hosts</filename>.</para>
990 <title>Testing and Troubleshooting</title>
992 <para>Once you have made the necessary changes in
993 <filename>/etc/rc.conf</filename>, you should reboot your
994 system. This will allow the change(s) to the interface(s) to
995 be applied, and verify that the system restarts without any
996 configuration errors.</para>
998 <para>Once the system has been rebooted, you should test the
999 network interfaces.</para>
1002 <title>Testing the Ethernet Card</title>
1005 <primary>network card configuration</primary>
1006 <secondary>testing the card</secondary>
1009 <para>To verify that an Ethernet card is configured correctly,
1010 you have to try two things. First, ping the interface itself,
1011 and then ping another machine on the LAN.</para>
1013 <para>First test the local interface:</para>
1015 <screen>&prompt.user; <userinput>ping -c5 192.168.1.3</userinput>
1016 PING 192.168.1.3 (192.168.1.3): 56 data bytes
1017 64 bytes from 192.168.1.3: icmp_seq=0 ttl=64 time=0.082 ms
1018 64 bytes from 192.168.1.3: icmp_seq=1 ttl=64 time=0.074 ms
1019 64 bytes from 192.168.1.3: icmp_seq=2 ttl=64 time=0.076 ms
1020 64 bytes from 192.168.1.3: icmp_seq=3 ttl=64 time=0.108 ms
1021 64 bytes from 192.168.1.3: icmp_seq=4 ttl=64 time=0.076 ms
1023 --- 192.168.1.3 ping statistics ---
1024 5 packets transmitted, 5 packets received, 0% packet loss
1025 round-trip min/avg/max/stddev = 0.074/0.083/0.108/0.013 ms</screen>
1027 <para>Now we have to ping another machine on the LAN:</para>
1029 <screen>&prompt.user; <userinput>ping -c5 192.168.1.2</userinput>
1030 PING 192.168.1.2 (192.168.1.2): 56 data bytes
1031 64 bytes from 192.168.1.2: icmp_seq=0 ttl=64 time=0.726 ms
1032 64 bytes from 192.168.1.2: icmp_seq=1 ttl=64 time=0.766 ms
1033 64 bytes from 192.168.1.2: icmp_seq=2 ttl=64 time=0.700 ms
1034 64 bytes from 192.168.1.2: icmp_seq=3 ttl=64 time=0.747 ms
1035 64 bytes from 192.168.1.2: icmp_seq=4 ttl=64 time=0.704 ms
1037 --- 192.168.1.2 ping statistics ---
1038 5 packets transmitted, 5 packets received, 0% packet loss
1039 round-trip min/avg/max/stddev = 0.700/0.729/0.766/0.025 ms</screen>
1041 <para>You could also use the machine name instead of
1042 <hostid role="ipaddr">192.168.1.2</hostid> if you have set up the
1043 <filename>/etc/hosts</filename> file.</para>
1047 <title>Troubleshooting</title>
1050 <primary>network card configuration</primary>
1051 <secondary>troubleshooting</secondary>
1054 <para>Troubleshooting hardware and software configurations is always
1055 a pain, and a pain which can be alleviated by checking the simple
1056 things first. Is your network cable plugged in? Have you properly
1057 configured the network services? Did you configure the firewall
1058 correctly? Is the card you are using supported by &os;? Always
1059 check the hardware notes before sending off a bug report. Update
1060 your version of &os; to the latest STABLE version. Check the
1061 mailing list archives, or perhaps search the Internet.</para>
1063 <para>If the card works, yet performance is poor, it would be
1064 worthwhile to read over the &man.tuning.7; manual page. You
1065 can also check the network configuration as incorrect network
1066 settings can cause slow connections.</para>
1068 <para>Some users experience one or two <quote>device
1069 timeouts</quote>, which is normal for some cards. If they
1070 continue, or are bothersome, you may wish to be sure the
1071 device is not conflicting with another device. Double check
1072 the cable connections. Perhaps you may just need to get
1073 another card.</para>
1075 <para>At times, users see a few <errorname>watchdog timeout</errorname>
1076 errors. The first thing to do here is to check your network
1077 cable. Many cards require a PCI slot which supports Bus
1078 Mastering. On some old motherboards, only one PCI slot allows
1079 it (usually slot 0). Check the network card and the
1080 motherboard documentation to determine if that may be the
1083 <para><errorname>No route to host</errorname> messages occur if the
1084 system is unable to route a packet to the destination host.
1085 This can happen if no default route is specified, or if a
1086 cable is unplugged. Check the output of <command>netstat
1087 -rn</command> and make sure there is a valid route to the host
1088 you are trying to reach. If there is not, read on to <xref
1089 linkend="advanced-networking">.</para>
1091 <para><errorname>ping: sendto: Permission denied</errorname> error
1092 messages are often caused by a misconfigured firewall. If
1093 <command>ipfw</command> is enabled in the kernel but no rules
1094 have been defined, then the default policy is to deny all
1095 traffic, even ping requests! Read on to <xref
1096 linkend="firewalls"> for more information.</para>
1098 <para>Sometimes performance of the card is poor, or below average.
1099 In these cases it is best to set the media selection mode
1100 from <literal>autoselect</literal> to the correct media selection.
1101 While this usually works for most hardware, it may not resolve
1102 this issue for everyone. Again, check all the network settings,
1103 and read over the &man.tuning.7; manual page.</para>
1109 <sect1 id="configtuning-virtual-hosts">
1110 <title>Virtual Hosts</title>
1112 <indexterm><primary>virtual hosts</primary></indexterm>
1113 <indexterm><primary>IP aliases</primary></indexterm>
1115 <para>A very common use of &os; is virtual site hosting, where
1116 one server appears to the network as many servers. This is
1117 achieved by assigning multiple network addresses to a single
1120 <para>A given network interface has one <quote>real</quote> address,
1121 and may have any number of <quote>alias</quote> addresses.
1123 normally added by placing alias entries in
1124 <filename>/etc/rc.conf</filename>.</para>
1126 <para>An alias entry for the interface <devicename>fxp0</devicename>
1129 <programlisting>ifconfig_fxp0_alias0="inet xxx.xxx.xxx.xxx netmask xxx.xxx.xxx.xxx"</programlisting>
1131 <para>Note that alias entries must start with alias0 and proceed
1132 upwards in order, (for example, _alias1, _alias2, and so on).
1133 The configuration process will stop at the first missing number.
1136 <para>The calculation of alias netmasks is important, but
1137 fortunately quite simple. For a given interface, there must be
1138 one address which correctly represents the network's netmask.
1139 Any other addresses which fall within this network must have a
1140 netmask of all <literal>1</literal>s.</para>
1142 <para>For example, consider the case where the
1143 <devicename>fxp0</devicename> interface is
1144 connected to two networks, the <hostid role="ipaddr">10.1.1.0</hostid>
1145 network with a netmask of <hostid role="netmask">255.255.255.0</hostid>
1146 and the <hostid role="ipaddr">202.0.75.16</hostid> network with
1147 a netmask of <hostid role="netmask">255.255.255.240</hostid>.
1148 We want the system to appear at <hostid role="ipaddr">10.1.1.1</hostid>
1149 through <hostid role="ipaddr">10.1.1.5</hostid> and at
1150 <hostid role="ipaddr">202.0.75.17</hostid> through
1151 <hostid role="ipaddr">202.0.75.20</hostid>.</para>
1153 <para>The following entries configure the adapter correctly for
1154 this arrangement:</para>
1156 <programlisting> ifconfig_fxp0="inet 10.1.1.1 netmask 255.255.255.0"
1157 ifconfig_fxp0_alias0="inet 10.1.1.2 netmask 255.255.255.255"
1158 ifconfig_fxp0_alias1="inet 10.1.1.3 netmask 255.255.255.255"
1159 ifconfig_fxp0_alias2="inet 10.1.1.4 netmask 255.255.255.255"
1160 ifconfig_fxp0_alias3="inet 10.1.1.5 netmask 255.255.255.255"
1161 ifconfig_fxp0_alias4="inet 202.0.75.17 netmask 255.255.255.240"
1162 ifconfig_fxp0_alias5="inet 202.0.75.18 netmask 255.255.255.255"
1163 ifconfig_fxp0_alias6="inet 202.0.75.19 netmask 255.255.255.255"
1164 ifconfig_fxp0_alias7="inet 202.0.75.20 netmask 255.255.255.255"</programlisting>
1168 <sect1 id="configtuning-configfiles">
1169 <title>Configuration Files</title>
1172 <title><filename>/etc</filename> Layout</title>
1173 <para>There are a number of directories in which configuration
1174 information is kept. These include:</para>
1176 <informaltable frame="none">
1180 <entry><filename>/etc</filename></entry>
1181 <entry>Generic system configuration information; data here is
1182 system-specific.</entry>
1185 <entry><filename>/etc/defaults</filename></entry>
1186 <entry>Default versions of system configuration files.</entry>
1189 <entry><filename>/etc/mail</filename></entry>
1190 <entry>Extra &man.sendmail.8; configuration, other
1191 MTA configuration files.
1195 <entry><filename>/etc/ppp</filename></entry>
1196 <entry>Configuration for both user- and kernel-ppp programs.
1200 <entry><filename>/etc/namedb</filename></entry>
1201 <entry>Default location for &man.named.8; data. Normally
1202 <filename>named.conf</filename> and zone files are stored
1206 <entry><filename>/usr/local/etc</filename></entry>
1207 <entry>Configuration files for installed applications.
1208 May contain per-application subdirectories.</entry>
1211 <entry><filename>/usr/local/etc/rc.d</filename></entry>
1212 <entry>Start/stop scripts for installed applications.</entry>
1215 <entry><filename>/var/db</filename></entry>
1216 <entry>Automatically generated system-specific database files,
1217 such as the package database, the locate database, and so
1226 <title>Hostnames</title>
1228 <indexterm><primary>hostname</primary></indexterm>
1229 <indexterm><primary>DNS</primary></indexterm>
1232 <title><filename>/etc/resolv.conf</filename></title>
1235 <primary><filename>resolv.conf</filename></primary>
1238 <para><filename>/etc/resolv.conf</filename> dictates how &os;'s
1239 resolver accesses the Internet Domain Name System (DNS).</para>
1241 <para>The most common entries to <filename>resolv.conf</filename> are:
1244 <informaltable frame="none">
1248 <entry><literal>nameserver</literal></entry>
1249 <entry>The IP address of a name server the resolver
1250 should query. The servers are queried in the order
1251 listed with a maximum of three.</entry>
1254 <entry><literal>search</literal></entry>
1255 <entry>Search list for hostname lookup. This is normally
1256 determined by the domain of the local hostname.</entry>
1259 <entry><literal>domain</literal></entry>
1260 <entry>The local domain name.</entry>
1266 <para>A typical <filename>resolv.conf</filename>:</para>
1268 <programlisting>search example.com
1269 nameserver 147.11.1.11
1270 nameserver 147.11.100.30</programlisting>
1272 <note><para>Only one of the <literal>search</literal> and
1273 <literal>domain</literal> options should be used.</para></note>
1275 <para>If you are using DHCP, &man.dhclient.8; usually rewrites
1276 <filename>resolv.conf</filename> with information received from the
1281 <title><filename>/etc/hosts</filename></title>
1283 <indexterm><primary>hosts</primary></indexterm>
1285 <para><filename>/etc/hosts</filename> is a simple text
1286 database reminiscent of the old Internet. It works in
1287 conjunction with DNS and NIS providing name to IP address
1288 mappings. Local computers connected via a LAN can be placed
1289 in here for simplistic naming purposes instead of setting up
1290 a &man.named.8; server. Additionally,
1291 <filename>/etc/hosts</filename> can be used to provide a
1292 local record of Internet names, reducing the need to query
1293 externally for commonly accessed names.</para>
1295 <programlisting># $&os;$
1298 # This file should contain the addresses and aliases
1299 # for local hosts that share this file.
1300 # In the presence of the domain name service or NIS, this file may
1301 # not be consulted at all; see /etc/nsswitch.conf for the resolution order.
1304 ::1 localhost localhost.my.domain myname.my.domain
1305 127.0.0.1 localhost localhost.my.domain myname.my.domain
1308 # Imaginary network.
1309 #10.0.0.2 myname.my.domain myname
1310 #10.0.0.3 myfriend.my.domain myfriend
1312 # According to RFC 1918, you can use the following IP networks for
1313 # private nets which will never be connected to the Internet:
1315 # 10.0.0.0 - 10.255.255.255
1316 # 172.16.0.0 - 172.31.255.255
1317 # 192.168.0.0 - 192.168.255.255
1319 # In case you want to be able to connect to the Internet, you need
1320 # real official assigned numbers. PLEASE PLEASE PLEASE do not try
1321 # to invent your own network numbers but instead get one from your
1322 # network provider (if any) or from the Internet Registry (ftp to
1323 # rs.internic.net, directory `/templates').
1326 <para><filename>/etc/hosts</filename> takes on the simple format
1329 <programlisting>[Internet address] [official hostname] [alias1] [alias2] ...</programlisting>
1331 <para>For example:</para>
1333 <programlisting>10.0.0.1 myRealHostname.example.com myRealHostname foobar1 foobar2</programlisting>
1335 <para>Consult &man.hosts.5; for more information.</para>
1340 <title>Log File Configuration</title>
1342 <indexterm><primary>log files</primary></indexterm>
1345 <title><filename>syslog.conf</filename></title>
1347 <indexterm><primary>syslog.conf</primary></indexterm>
1349 <para><filename>syslog.conf</filename> is the configuration file
1350 for the &man.syslogd.8; program. It indicates which types
1351 of <command>syslog</command> messages are logged to particular
1354 <programlisting># $&os;$
1356 # Spaces ARE valid field separators in this file. However,
1357 # other *nix-like systems still insist on using tabs as field
1358 # separators. If you are sharing this file between systems, you
1359 # may want to use only tabs as field separators here.
1360 # Consult the syslog.conf(5) manual page.
1361 *.err;kern.debug;auth.notice;mail.crit /dev/console
1362 *.notice;kern.debug;lpr.info;mail.crit;news.err /var/log/messages
1363 security.* /var/log/security
1364 mail.info /var/log/maillog
1365 lpr.info /var/log/lpd-errs
1366 cron.* /var/log/cron
1368 *.notice;news.err root
1371 # uncomment this to log all writes to /dev/console to /var/log/console.log
1372 #console.info /var/log/console.log
1373 # uncomment this to enable logging of all log messages to /var/log/all.log
1374 #*.* /var/log/all.log
1375 # uncomment this to enable logging to a remote log host named loghost
1377 # uncomment these if you're running inn
1378 # news.crit /var/log/news/news.crit
1379 # news.err /var/log/news/news.err
1380 # news.notice /var/log/news/news.notice
1382 *.* /var/log/slip.log
1384 *.* /var/log/ppp.log</programlisting>
1386 <para>Consult the &man.syslog.conf.5; manual page for more
1391 <title><filename>newsyslog.conf</filename></title>
1393 <indexterm><primary>newsyslog.conf</primary></indexterm>
1395 <para><filename>newsyslog.conf</filename> is the configuration
1396 file for &man.newsyslog.8;, a program that is normally scheduled
1397 to run by &man.cron.8;. &man.newsyslog.8; determines when log
1398 files require archiving or rearranging.
1399 <filename>logfile</filename> is moved to
1400 <filename>logfile.0</filename>, <filename>logfile.0</filename>
1401 is moved to <filename>logfile.1</filename>, and so on.
1402 Alternatively, the log files may be archived in &man.gzip.1; format
1403 causing them to be named: <filename>logfile.0.gz</filename>,
1404 <filename>logfile.1.gz</filename>, and so on.</para>
1406 <para><filename>newsyslog.conf</filename> indicates which log
1407 files are to be managed, how many are to be kept, and when
1408 they are to be touched. Log files can be rearranged and/or
1409 archived when they have either reached a certain size, or at a
1410 certain periodic time/date.</para>
1412 <programlisting># configuration file for newsyslog
1413 # $&os;$
1415 # filename [owner:group] mode count size when [ZB] [/pid_file] [sig_num]
1416 /var/log/cron 600 3 100 * Z
1417 /var/log/amd.log 644 7 100 * Z
1418 /var/log/kerberos.log 644 7 100 * Z
1419 /var/log/lpd-errs 644 7 100 * Z
1420 /var/log/maillog 644 7 * @T00 Z
1421 /var/log/sendmail.st 644 10 * 168 B
1422 /var/log/messages 644 5 100 * Z
1423 /var/log/all.log 600 7 * @T00 Z
1424 /var/log/slip.log 600 3 100 * Z
1425 /var/log/ppp.log 600 3 100 * Z
1426 /var/log/security 600 10 100 * Z
1427 /var/log/wtmp 644 3 * @01T05 B
1428 /var/log/daily.log 640 7 * @T00 Z
1429 /var/log/weekly.log 640 5 1 $W6D0 Z
1430 /var/log/monthly.log 640 12 * $M1D0 Z
1431 /var/log/console.log 640 5 100 * Z</programlisting>
1433 <para>Consult the &man.newsyslog.8; manual page for more
1439 <title><filename>sysctl.conf</filename></title>
1441 <indexterm><primary>sysctl.conf</primary></indexterm>
1442 <indexterm><primary>sysctl</primary></indexterm>
1444 <para><filename>sysctl.conf</filename> looks much like
1445 <filename>rc.conf</filename>. Values are set in a
1446 <literal>variable=value</literal>
1447 form. The specified values are set after the system goes into
1448 multi-user mode. Not all variables are settable in this mode.</para>
1450 <para>A sample <filename>sysctl.conf</filename> turning off logging
1451 of fatal signal exits and letting Linux programs know they are really
1452 running under &os;:</para>
1454 <programlisting>kern.logsigexit=0 # Do not log fatal signal exits (e.g. sig 11)
1455 compat.linux.osname=&os;
1456 compat.linux.osrelease=4.3-STABLE</programlisting>
1460 <sect1 id="configtuning-sysctl">
1461 <title>Tuning with sysctl</title>
1463 <indexterm><primary>sysctl</primary></indexterm>
1465 <primary>tuning</primary>
1466 <secondary>with sysctl</secondary>
1469 <para>&man.sysctl.8; is an interface that allows you to make changes
1470 to a running &os; system. This includes many advanced
1471 options of the TCP/IP stack and virtual memory system that can
1472 dramatically improve performance for an experienced system
1473 administrator. Over five hundred system variables can be read
1474 and set using &man.sysctl.8;.</para>
1476 <para>At its core, &man.sysctl.8; serves two functions: to read and
1477 to modify system settings.</para>
1479 <para>To view all readable variables:</para>
1481 <screen>&prompt.user; <userinput>sysctl -a</userinput></screen>
1483 <para>To read a particular variable, for example,
1484 <varname>kern.maxproc</varname>:</para>
1486 <screen>&prompt.user; <userinput>sysctl kern.maxproc</userinput>
1487 kern.maxproc: 1044</screen>
1489 <para>To set a particular variable, use the intuitive
1490 <replaceable>variable</replaceable>=<replaceable>value</replaceable>
1493 <screen>&prompt.root; <userinput>sysctl kern.maxfiles=5000</userinput>
1494 kern.maxfiles: 2088 -> 5000</screen>
1496 <para>Settings of sysctl variables are usually either strings,
1497 numbers, or booleans (a boolean being <literal>1</literal> for yes
1498 or a <literal>0</literal> for no).</para>
1500 <sect2 id="sysctl-readonly">
1504 <firstname>Tom</firstname>
1505 <surname>Rhodes</surname>
1506 <contrib>Contributed by </contrib>
1507 <!-- 31 January 2003 -->
1511 <title>&man.sysctl.8; Read-only</title>
1513 <para>In some cases it may be desirable to modify read-only &man.sysctl.8;
1514 values. While this is not recommended, it is also sometimes unavoidable.</para>
1516 <para>For instance on some laptop models the &man.cardbus.4; device will
1517 not probe memory ranges, and fail with errors which look similar to:</para>
1519 <screen>cbb0: Could not map register memory
1520 device_probe_and_attach: cbb0 attach returned 12</screen>
1522 <para>Cases like the one above usually require the modification of some
1523 default &man.sysctl.8; settings which are set read only. To overcome
1524 these situations a user can put &man.sysctl.8; <quote>OIDs</quote>
1525 in their local <filename>/boot/loader.conf</filename>. Default
1526 settings are located in the <filename>/boot/defaults/loader.conf</filename>
1529 <para>Fixing the problem mentioned above would require a user to set
1530 <option>hw.pci.allow_unsupported_io_range=1</option> in the aforementioned
1531 file. Now &man.cardbus.4; will work properly.</para>
1536 <sect1 id="configtuning-disk">
1537 <title>Tuning Disks</title>
1540 <title>Sysctl Variables</title>
1543 <title><varname>vfs.vmiodirenable</varname></title>
1546 <primary><varname>vfs.vmiodirenable</varname></primary>
1549 <para>The <varname>vfs.vmiodirenable</varname> sysctl variable
1550 may be set to either 0 (off) or 1 (on); it is 1 by default.
1551 This variable controls how directories are cached by the
1552 system. Most directories are small, using just a single
1553 fragment (typically 1 K) in the file system and less
1554 (typically 512 bytes) in the buffer cache.
1555 However, when operating in the default mode the buffer
1556 cache will only cache a fixed number of directories even if
1557 you have a huge amount of memory. Turning on this sysctl
1558 allows the buffer cache to use the VM Page Cache to cache the
1559 directories, making all the memory available for caching
1560 directories. However,
1561 the minimum in-core memory used to cache a directory is the
1562 physical page size (typically 4 K) rather than 512
1563 bytes. We recommend turning this option on if you are running
1564 any services which manipulate large numbers of files. Such
1565 services can include web caches, large mail systems, and news
1566 systems. Turning on this option will generally not reduce
1567 performance even with the wasted memory but you should
1568 experiment to find out.</para>
1572 <title><varname>vfs.write_behind</varname></title>
1575 <primary><varname>vfs.write_behind</varname></primary>
1578 <para>The <varname>vfs.write_behind</varname> sysctl variable
1579 defaults to <literal>1</literal> (on). This tells the file system
1580 to issue media writes as full clusters are collected, which
1581 typically occurs when writing large sequential files. The idea
1582 is to avoid saturating the buffer cache with dirty buffers when
1583 it would not benefit I/O performance. However, this may stall
1584 processes and under certain circumstances you may wish to turn it
1589 <title><varname>vfs.hirunningspace</varname></title>
1592 <primary><varname>vfs.hirunningspace</varname></primary>
1595 <para>The <varname>vfs.hirunningspace</varname> sysctl variable
1596 determines how much outstanding write I/O may be queued to disk
1597 controllers system-wide at any given instance. The default is
1598 usually sufficient but on machines with lots of disks you may
1599 want to bump it up to four or five <emphasis>megabytes</emphasis>.
1600 Note that setting too high a value (exceeding the buffer cache's
1601 write threshold) can lead to extremely bad clustering
1602 performance. Do not set this value arbitrarily high! Higher
1603 write values may add latency to reads occurring at the same time.
1606 <para>There are various other buffer-cache and VM page cache
1607 related sysctls. We do not recommend modifying these values.
1608 The VM system does an extremely good job of
1609 automatically tuning itself.</para>
1613 <title><varname>vm.swap_idle_enabled</varname></title>
1616 <primary><varname>vm.swap_idle_enabled</varname></primary>
1619 <para>The <varname>vm.swap_idle_enabled</varname> sysctl variable
1620 is useful in large multi-user systems where you have lots of
1621 users entering and leaving the system and lots of idle processes.
1622 Such systems tend to generate a great deal of continuous pressure
1623 on free memory reserves. Turning this feature on and tweaking
1624 the swapout hysteresis (in idle seconds) via
1625 <varname>vm.swap_idle_threshold1</varname> and
1626 <varname>vm.swap_idle_threshold2</varname> allows you to depress
1627 the priority of memory pages associated with idle processes more
1628 quickly then the normal pageout algorithm. This gives a helping
1629 hand to the pageout daemon. Do not turn this option on unless
1630 you need it, because the tradeoff you are making is essentially
1631 pre-page memory sooner rather than later; thus eating more swap
1632 and disk bandwidth. In a small system this option will have a
1633 determinable effect but in a large system that is already doing
1634 moderate paging this option allows the VM system to stage whole
1635 processes into and out of memory easily.</para>
1639 <title><varname>hw.ata.wc</varname></title>
1642 <primary><varname>hw.ata.wc</varname></primary>
1645 <para>IDE drives lie about when a write completes. With IDE write
1646 caching turned on, IDE hard drives not only write data
1647 to disk out of order, but will sometimes delay writing some
1648 blocks indefinitely when under heavy disk loads. A crash or
1649 power failure may cause serious file system corruption. Turning
1650 off write caching will remove the danger of this data loss, but
1651 will also cause disk operations to proceed
1652 <emphasis>very slowly.</emphasis> Change this only if prepared
1653 to suffer with the disk slowdown.</para>
1655 <para>Changing this variable must be done from the
1656 boot loader at boot time. Attempting to do it after the
1657 kernel boots will have no effect.</para>
1659 <para>For more information, please see &man.ata.4;.</para>
1664 <sect2 id="soft-updates">
1665 <title>Soft Updates</title>
1667 <indexterm><primary>Soft Updates</primary></indexterm>
1668 <indexterm><primary>tunefs</primary></indexterm>
1670 <para>The &man.tunefs.8; program can be used to fine-tune a
1671 file system. This program has many different options, but for
1672 now we are only concerned with toggling Soft Updates on and
1673 off, which is done by:</para>
1675 <screen>&prompt.root; <userinput>tunefs -n enable /filesystem</userinput>
1676 &prompt.root; <userinput>tunefs -n disable /filesystem</userinput></screen>
1678 <para>A filesystem cannot be modified with &man.tunefs.8; while
1679 it is mounted. A good time to enable Soft Updates is before any
1680 partitions have been mounted, in single-user mode.</para>
1682 <note><para>It is possible to enable Soft Updates
1683 at filesystem creation time, through use of the <literal>-U</literal>
1684 option to &man.newfs.8;.</para></note>
1686 <para>Soft Updates drastically improves meta-data performance, mainly
1687 file creation and deletion, through the use of a memory cache. We
1688 recommend to use Soft Updates on all of your file systems. There
1689 are two downsides to Soft Updates that you should be aware of: First,
1690 Soft Updates guarantees filesystem consistency in the case of a crash
1691 but could very easily be several seconds (even a minute!) behind
1692 updating the physical disk. If your system crashes you may lose more
1693 work than otherwise. Secondly, Soft Updates delays the freeing of
1694 filesystem blocks. If you have a filesystem (such as the root
1695 filesystem) which is almost full, performing a major update, such as
1696 <command>make installworld</command>, can cause the filesystem to run
1697 out of space and the update to fail.</para>
1700 <title>More Details about Soft Updates</title>
1703 <primary>Soft Updates</primary>
1704 <secondary>details</secondary>
1707 <para>There are two traditional approaches to writing a file
1708 systems meta-data back to disk. (Meta-data updates are
1709 updates to non-content data like inodes or
1710 directories.)</para>
1712 <para>Historically, the default behavior was to write out
1713 meta-data updates synchronously. If a directory had been
1714 changed, the system waited until the change was actually
1715 written to disk. The file data buffers (file contents) were
1716 passed through the buffer cache and backed up
1717 to disk later on asynchronously. The advantage of this
1718 implementation is that it operates safely. If there is
1719 a failure during an update, the meta-data are always in a
1720 consistent state. A file is either created completely
1721 or not at all. If the data blocks of a file did not find
1722 their way out of the buffer cache onto the disk by the time
1723 of the crash, &man.fsck.8; is able to recognize this and
1724 repair the filesystem by setting the file length to
1725 0. Additionally, the implementation is clear and simple.
1726 The disadvantage is that meta-data changes are slow. An
1727 <command>rm -r</command>, for instance, touches all the files
1728 in a directory sequentially, but each directory
1729 change (deletion of a file) will be written synchronously
1730 to the disk. This includes updates to the directory itself,
1731 to the inode table, and possibly to indirect blocks
1732 allocated by the file. Similar considerations apply for
1733 unrolling large hierarchies (<command>tar -x</command>).</para>
1735 <para>The second case is asynchronous meta-data updates. This
1736 is the default for Linux/ext2fs and
1737 <command>mount -o async</command> for *BSD ufs. All
1738 meta-data updates are simply being passed through the buffer
1739 cache too, that is, they will be intermixed with the updates
1740 of the file content data. The advantage of this
1741 implementation is there is no need to wait until each
1742 meta-data update has been written to disk, so all operations
1743 which cause huge amounts of meta-data updates work much
1744 faster than in the synchronous case. Also, the
1745 implementation is still clear and simple, so there is a low
1746 risk for bugs creeping into the code. The disadvantage is
1747 that there is no guarantee at all for a consistent state of
1748 the filesystem. If there is a failure during an operation
1749 that updated large amounts of meta-data (like a power
1750 failure, or someone pressing the reset button),
1752 will be left in an unpredictable state. There is no opportunity
1753 to examine the state of the filesystem when the system
1754 comes up again; the data blocks of a file could already have
1755 been written to the disk while the updates of the inode
1756 table or the associated directory were not. It is actually
1757 impossible to implement a <command>fsck</command> which is
1758 able to clean up the resulting chaos (because the necessary
1759 information is not available on the disk). If the
1760 filesystem has been damaged beyond repair, the only choice
1761 is to use &man.newfs.8; on it and restore it from backup.
1764 <para>The usual solution for this problem was to implement
1765 <emphasis>dirty region logging</emphasis>, which is also
1766 referred to as <emphasis>journaling</emphasis>, although that
1767 term is not used consistently and is occasionally applied
1768 to other forms of transaction logging as well. Meta-data
1769 updates are still written synchronously, but only into a
1770 small region of the disk. Later on they will be moved
1771 to their proper location. Because the logging
1772 area is a small, contiguous region on the disk, there
1773 are no long distances for the disk heads to move, even
1774 during heavy operations, so these operations are quicker
1775 than synchronous updates.
1776 Additionally the complexity of the implementation is fairly
1777 limited, so the risk of bugs being present is low. A disadvantage
1778 is that all meta-data are written twice (once into the
1779 logging region and once to the proper location) so for
1780 normal work, a performance <quote>pessimization</quote>
1781 might result. On the other hand, in case of a crash, all
1782 pending meta-data operations can be quickly either rolled-back
1783 or completed from the logging area after the system comes
1784 up again, resulting in a fast filesystem startup.</para>
1786 <para>Kirk McKusick, the developer of Berkeley FFS,
1787 solved this problem with Soft Updates: all pending
1788 meta-data updates are kept in memory and written out to disk
1789 in a sorted sequence (<quote>ordered meta-data
1790 updates</quote>). This has the effect that, in case of
1791 heavy meta-data operations, later updates to an item
1792 <quote>catch</quote> the earlier ones if the earlier ones are still in
1793 memory and have not already been written to disk. So all
1794 operations on, say, a directory are generally performed in
1795 memory before the update is written to disk (the data
1796 blocks are sorted according to their position so
1797 that they will not be on the disk ahead of their meta-data).
1798 If the system crashes, this causes an implicit <quote>log
1799 rewind</quote>: all operations which did not find their way
1800 to the disk appear as if they had never happened. A
1801 consistent filesystem state is maintained that appears to
1802 be the one of 30 to 60 seconds earlier. The
1803 algorithm used guarantees that all resources in use
1804 are marked as such in their appropriate bitmaps: blocks and inodes.
1805 After a crash, the only resource allocation error
1806 that occurs is that resources are
1807 marked as <quote>used</quote> which are actually <quote>free</quote>.
1808 &man.fsck.8; recognizes this situation,
1809 and frees the resources that are no longer used. It is safe to
1810 ignore the dirty state of the filesystem after a crash by
1811 forcibly mounting it with <command>mount -f</command>. In
1812 order to free resources that may be unused, &man.fsck.8;
1813 needs to be run at a later time.</para>
1815 <para>The advantage is that meta-data operations are nearly as
1816 fast as asynchronous updates (i.e. faster than with
1817 <emphasis>logging</emphasis>, which has to write the
1818 meta-data twice). The disadvantages are the complexity of
1819 the code (implying a higher risk for bugs in an area that
1820 is highly sensitive regarding loss of user data), and a
1821 higher memory consumption. Additionally there are some
1822 idiosyncrasies one has to get used to.
1823 After a crash, the state of the filesystem appears to be
1824 somewhat <quote>older</quote>. In situations where
1825 the standard synchronous approach would have caused some
1826 zero-length files to remain after the
1827 <command>fsck</command>, these files do not exist at all
1828 with a Soft Updates filesystem because neither the meta-data
1829 nor the file contents have ever been written to disk.
1830 Disk space is not released until the updates have been
1831 written to disk, which may take place some time after
1832 running <command>rm</command>. This may cause problems
1833 when installing large amounts of data on a filesystem
1834 that does not have enough free space to hold all the files
1840 <sect1 id="configtuning-kernel-limits">
1841 <title>Tuning Kernel Limits</title>
1844 <primary>tuning</primary>
1845 <secondary>kernel limits</secondary>
1848 <sect2 id="file-process-limits">
1849 <title>File/Process Limits</title>
1851 <sect3 id="kern-maxfiles">
1852 <title><varname>kern.maxfiles</varname></title>
1855 <primary><varname>kern.maxfiles</varname></primary>
1858 <para><varname>kern.maxfiles</varname> can be raised or
1859 lowered based upon your system requirements. This variable
1860 indicates the maximum number of file descriptors on your
1861 system. When the file descriptor table is full,
1862 <errorname>file: table is full</errorname> will show up repeatedly
1863 in the system message buffer, which can be viewed with the
1864 <command>dmesg</command> command.</para>
1866 <para>Each open file, socket, or fifo uses one file
1867 descriptor. A large-scale production server may easily
1868 require many thousands of file descriptors, depending on the
1869 kind and number of services running concurrently.</para>
1871 <para><varname>kern.maxfile</varname>'s default value is
1872 dictated by the <option>MAXUSERS</option> option in your
1873 kernel configuration file. <varname>kern.maxfiles</varname> grows
1874 proportionally to the value of <option>MAXUSERS</option>. When
1875 compiling a custom kernel, it is a good idea to set this kernel
1876 configuration option according to the uses of your system. From
1877 this number, the kernel is given most of its pre-defined limits.
1878 Even though a production machine may not actually have 256 users
1879 connected at once, the resources needed may be similar to a
1880 high-scale web server.</para>
1882 <note><para>Setting <option>MAXUSERS</option> to
1883 <literal>0</literal> in your kernel configuration file will choose
1884 a reasonable default value based on the amount of RAM present in
1885 your system. It is set to 0 in the default GENERIC kernel.</para></note>
1890 <title><varname>kern.ipc.somaxconn</varname></title>
1893 <primary><varname>kern.ipc.somaxconn</varname></primary>
1896 <para>The <varname>kern.ipc.somaxconn</varname> sysctl variable
1897 limits the size of the listen queue for accepting new TCP
1898 connections. The default value of <literal>128</literal> is
1899 typically too low for robust handling of new connections in a
1900 heavily loaded web server environment. For such environments, it
1901 is recommended to increase this value to <literal>1024</literal> or
1902 higher. The service daemon may itself limit the listen queue size
1903 (e.g. &man.sendmail.8;, or <application>Apache</application>) but
1904 will often have a directive in its configuration file to adjust
1905 the queue size. Large listen queues also do a better job of
1906 avoiding Denial of Service (<abbrev>DoS</abbrev>) attacks.</para>
1911 <title>Network Limits</title>
1913 <para>The <option>NMBCLUSTERS</option> kernel configuration
1914 option dictates the amount of network Mbufs available to the
1915 system. A heavily-trafficked server with a low number of Mbufs
1916 will hinder &os;'s ability. Each cluster represents
1917 approximately 2 K of memory, so a value of 1024 represents 2
1918 megabytes of kernel memory reserved for network buffers. A
1919 simple calculation can be done to figure out how many are
1920 needed. If you have a web server which maxes out at 1000
1921 simultaneous connections, and each connection eats a 16 K receive
1922 and 16 K send buffer, you need approximately 32 MB worth of
1923 network buffers to cover the web server. A good rule of thumb is
1924 to multiply by 2, so 2x32 MB / 2 KB =
1925 64 MB / 2 kB = 32768. We recommend
1926 values between 4096 and 32768 for machines with greater amounts
1927 of memory. Under no circumstances should you specify an
1928 arbitrarily high value for this parameter as it could lead to a
1929 boot time crash. The <option>-m</option> option to
1930 &man.netstat.1; may be used to observe network cluster
1931 use. <varname>kern.ipc.nmbclusters</varname> loader tunable should
1932 be used to tune this at boot time.</para>
1934 <para>For busy servers that make extensive use of the
1935 &man.sendfile.2; system call, it may be necessary to increase
1936 the number of &man.sendfile.2; buffers via the
1937 <option>NSFBUFS</option> kernel configuration option or by
1938 setting its value in <filename>/boot/loader.conf</filename>
1939 (see &man.loader.8; for details). A common indicator that
1940 this parameter needs to be adjusted is when processes are seen
1941 in the <errorname>sfbufa</errorname> state. The sysctl
1942 variable <varname>kern.ipc.nsfbufs</varname> is a read-only
1943 glimpse at the kernel configured variable. This parameter
1944 nominally scales with <varname>kern.maxusers</varname>,
1945 however it may be necessary to tune accordingly.</para>
1948 <para>Even though a socket has been marked as non-blocking,
1949 calling &man.sendfile.2; on the non-blocking socket may
1950 result in the &man.sendfile.2; call blocking until enough
1951 <literal>struct sf_buf</literal>'s are made
1956 <title><varname>net.inet.ip.portrange.*</varname></title>
1959 <primary>net.inet.ip.portrange.*</primary>
1962 <para>The <varname>net.inet.ip.portrange.*</varname> sysctl
1963 variables control the port number ranges automatically bound to TCP
1964 and UDP sockets. There are three ranges: a low range, a default
1965 range, and a high range. Most network programs use the default
1966 range which is controlled by the
1967 <varname>net.inet.ip.portrange.first</varname> and
1968 <varname>net.inet.ip.portrange.last</varname>, which default to
1969 1024 and 5000, respectively. Bound port ranges are used for
1970 outgoing connections, and it is possible to run the system out of
1971 ports under certain circumstances. This most commonly occurs
1972 when you are running a heavily loaded web proxy. The port range
1973 is not an issue when running servers which handle mainly incoming
1974 connections, such as a normal web server, or has a limited number
1975 of outgoing connections, such as a mail relay. For situations
1976 where you may run yourself out of ports, it is recommended to
1977 increase <varname>net.inet.ip.portrange.last</varname> modestly.
1978 A value of <literal>10000</literal>, <literal>20000</literal> or
1979 <literal>30000</literal> may be reasonable. You should also
1980 consider firewall effects when changing the port range. Some
1981 firewalls may block large ranges of ports (usually low-numbered
1982 ports) and expect systems to use higher ranges of ports for
1983 outgoing connections — for this reason it is recommended that
1984 <varname>net.inet.ip.portrange.first</varname> be lowered.</para>
1988 <title>TCP Bandwidth Delay Product</title>
1991 <primary>TCP Bandwidth Delay Product Limiting</primary>
1992 <secondary><varname>net.inet.tcp.inflight_enable</varname></secondary>
1995 <para>The TCP Bandwidth Delay Product Limiting is similar to
1996 TCP/Vegas in <application>NetBSD</application>. It can be
1997 enabled by setting <varname>net.inet.tcp.inflight_enable</varname>
1998 sysctl variable to <literal>1</literal>. The system will attempt
1999 to calculate the bandwidth delay product for each connection and
2000 limit the amount of data queued to the network to just the amount
2001 required to maintain optimum throughput.</para>
2003 <para>This feature is useful if you are serving data over modems,
2004 Gigabit Ethernet, or even high speed WAN links (or any other link
2005 with a high bandwidth delay product), especially if you are also
2006 using window scaling or have configured a large send window. If
2007 you enable this option, you should also be sure to set
2008 <varname>net.inet.tcp.inflight_debug</varname> to
2009 <literal>0</literal> (disable debugging), and for production use
2010 setting <varname>net.inet.tcp.inflight_min</varname> to at least
2011 <literal>6144</literal> may be beneficial. However, note that
2012 setting high minimums may effectively disable bandwidth limiting
2013 depending on the link. The limiting feature reduces the amount of
2014 data built up in intermediate route and switch packet queues as
2015 well as reduces the amount of data built up in the local host's
2016 interface queue. With fewer packets queued up, interactive
2017 connections, especially over slow modems, will also be able to
2018 operate with lower <emphasis>Round Trip Times</emphasis>. However,
2019 note that this feature only effects data transmission (uploading
2020 / server side). It has no effect on data reception (downloading).
2023 <para>Adjusting <varname>net.inet.tcp.inflight_stab</varname> is
2024 <emphasis>not</emphasis> recommended. This parameter defaults to
2025 20, representing 2 maximal packets added to the bandwidth delay
2026 product window calculation. The additional window is required to
2027 stabilize the algorithm and improve responsiveness to changing
2028 conditions, but it can also result in higher ping times over slow
2029 links (though still much lower than you would get without the
2030 inflight algorithm). In such cases, you may wish to try reducing
2031 this parameter to 15, 10, or 5; and may also have to reduce
2032 <varname>net.inet.tcp.inflight_min</varname> (for example, to
2033 3500) to get the desired effect. Reducing these parameters
2034 should be done as a last resort only.</para>
2039 <sect1 id="adding-swap-space">
2040 <title>Adding Swap Space</title>
2042 <para>No matter how well you plan, sometimes a system does not run
2043 as you expect. If you find you need more swap space, it is
2044 simple enough to add. You have three ways to increase swap
2045 space: adding a new hard drive, enabling swap over NFS, and
2046 creating a swap file on an existing partition.</para>
2048 <sect2 id="new-drive-swap">
2049 <title>Swap on a New Hard Drive</title>
2051 <para>The best way to add swap, of course, is to use this as an
2052 excuse to add another hard drive. You can always use another
2053 hard drive, after all. If you can do this, go reread the
2054 discussion of <ulink
2055 url="configtuning-initial.html#SWAP-DESIGN">swap space
2056 </ulink> from the <ulink
2057 url="configtuning-initial.html">Initial Configuration</ulink>
2058 section of the Handbook for some suggestions on how to best
2059 arrange your swap.</para>
2062 <sect2 id="nfs-swap">
2063 <title>Swapping over NFS</title>
2065 <para>Swapping over NFS is only recommended if you do not have a
2066 local hard disk to swap to. Even though &os; has an excellent
2067 NFS implementation, NFS swapping will be limited
2068 by the available network bandwidth and puts an additional
2069 burden on the NFS server.</para>
2072 <sect2 id="create-swapfile">
2073 <title>Swapfiles</title>
2075 <para>You can create a file of a specified size to use as a swap
2076 file. In our example here we will use a 64MB file called
2077 <filename>/usr/swap0</filename>. You can use any name you
2078 want, of course.</para>
2081 <title>Creating a Swapfile</title>
2085 <para>Be certain that your kernel configuration includes
2086 the vnode driver. It is <emphasis>not</emphasis> in recent versions of
2087 <filename>GENERIC</filename>.</para>
2089 <programlisting>pseudo-device vn 1 #Vnode driver (turns a file into a device)</programlisting>
2093 <para>Create a vn-device:</para>
2094 <screen>&prompt.root; <userinput>cd /dev</userinput>
2095 &prompt.root; <userinput>sh MAKEDEV vn0</userinput></screen>
2099 <para>Create a swapfile (<filename>/usr/swap0</filename>):</para>
2101 <screen>&prompt.root; <userinput>dd if=/dev/zero of=/usr/swap0 bs=1024k count=64</userinput></screen>
2105 <para>Set proper permissions on (<filename>/usr/swap0</filename>):</para>
2107 <screen>&prompt.root; <userinput>chmod 0600 /usr/swap0</userinput></screen>
2111 <para>Enable the swap file in <filename>/etc/rc.conf</filename>:</para>
2113 <programlisting>swapfile="/usr/swap0" # Set to name of swapfile if aux swapfile desired.</programlisting>
2118 <para>Reboot the machine or to enable the swap file immediately,
2121 <screen>&prompt.root; <userinput>vnconfig -e /dev/vn0b /usr/swap0 swap</userinput></screen>
2129 <sect1 id="acpi-overview">
2133 <firstname>Hiten</firstname>
2134 <surname>Pandya</surname>
2135 <contrib>Written by </contrib>
2138 <firstname>Tom</firstname>
2139 <surname>Rhodes</surname>
2144 <title>Power and Resource Management</title>
2146 <para>It is very important to utilize hardware resources in an
2147 efficient manner. Before <acronym>ACPI</acronym> was introduced,
2148 it was very difficult and inflexible for operating systems to manage
2149 the power usage and thermal properties of a system. The hardware was
2150 controlled by some sort of <acronym>BIOS</acronym> embedded
2151 interface, such as <emphasis>Plug and Play BIOS (PNPBIOS)</emphasis>, or
2152 <emphasis>Advanced Power Management (APM)</emphasis> and so on.
2153 Power and Resource Management is one of the key components of a modern
2154 operating system. For example, you may want an operating system to
2155 monitor system limits (and possibly alert you) in case your system
2156 temperature increased unexpectedly.</para>
2158 <para>In this section, we will provide
2159 comprehensive information about <acronym>ACPI</acronym>. References
2160 will be provided for further reading at the end. Please be aware
2161 that <acronym>ACPI</acronym> is available on &os; systems as a
2162 default kernel module. </para>
2164 <sect2 id="acpi-intro">
2165 <title>What Is ACPI?</title>
2167 <para>Advanced Configuration and Power Interface
2168 (<acronym>ACPI</acronym>) is a standard written by
2169 an alliance of vendors to provide a standard interface for
2170 hardware resources and power management (hence the name).
2171 It is a key element in <emphasis>Operating System-directed
2172 configuration and Power Management</emphasis>, i.e.: it provides
2173 more control and flexibility to the operating system
2174 (<acronym>OS</acronym>).
2175 Modern systems <quote>stretched</quote> the limits of the
2176 current Plug and Play interfaces (such as APM), prior to the introduction of
2177 <acronym>ACPI</acronym>. <acronym>ACPI</acronym> is the direct
2178 successor to <acronym>APM</acronym>
2179 (Advanced Power Management).</para>
2182 <sect2 id="acpi-old-spec">
2183 <title>Shortcomings of Advanced Power Management (APM)</title>
2185 <para>The <emphasis>Advanced Power Management (APM)</emphasis>
2186 facility control's the power usage of a system based on its
2187 activity. The APM BIOS is supplied by the (system) vendor and
2188 it is specific to the hardware platform. An APM driver in the
2189 OS mediates access to the <emphasis>APM Software Interface</emphasis>,
2190 which allows management of power levels.</para>
2192 <para>There are four major problems in APM. Firstly, power
2193 management is done by the (vendor-specific) BIOS, and the OS
2194 does not have any knowledge of it. One example of this, is when
2195 the user sets idle-time values for a hard drive in the APM BIOS,
2196 that when exceeded, it (BIOS) would spin down the hard drive,
2197 without the consent of the OS. Secondly, the APM logic is
2198 embedded in the BIOS, and it operates outside the scope of the
2199 OS. This means users can only fix problems in their APM BIOS by
2200 flashing a new one into the ROM; which, is a very dangerous
2201 procedure, and if it fails, it could leave the system in an
2202 unrecoverable state. Thirdly, APM is a vendor-specific
2203 technology, which, means that there is a lot or parity
2204 (duplication of efforts) and bugs found in one vendor's BIOS,
2205 may not be solved in others. Last but not the least, the APM
2206 BIOS did not have enough room to implement a sophisticated power
2207 policy, or one that can adapt very well to the purpose of the
2210 <para><emphasis>Plug and Play BIOS (PNPBIOS)</emphasis> was
2211 unreliable in many situations. PNPBIOS is 16-bit technology,
2212 so the OS has to use 16-bit emulation in order to
2213 <quote>interface</quote> with PNPBIOS methods.</para>
2215 <para>The &os; <acronym>APM</acronym> driver is documented in
2216 the &man.apm.4; manual page.</para>
2219 <sect2 id="acpi-config">
2220 <title>Configuring <acronym>ACPI</acronym></title>
2222 <para>The <filename>acpi.ko</filename> driver is loaded by default
2223 at start up by the &man.loader.8; and should <emphasis>not</emphasis>
2224 be compiled into the kernel. The reasoning behind this is that modules
2225 are easier to work with, say if switching to another <filename>acpi.ko</filename>
2226 without doing a kernel rebuild. This has the advantage of making testing easier.
2227 Another reason is that starting <acronym>ACPI</acronym> after a system has been
2228 brought up is not too useful, and in some cases can be fatal. In doubt, just
2229 disable <acronym>ACPI</acronym> all together. This driver should not and can not
2230 be unloaded because the system bus uses it for various hardware interactions.
2231 <acronym>ACPI</acronym> can be disabled with the &man.acpiconf.8; utility.
2232 In fact most of the interaction with <acronym>ACPI</acronym> can be done via
2233 &man.acpiconf.8;. Basically this means, if anything about <acronym>ACPI</acronym>
2234 is in the &man.dmesg.8; output, then most likely it is already running.</para>
2236 <note><para><acronym>ACPI</acronym> and <acronym>APM</acronym> cannot coexist and
2237 should be used separately. The last one to load will terminate if the driver
2238 notices the other running.</para></note>
2240 <para>In the simplest form, <acronym>ACPI</acronym> can be used to put the
2241 system into a sleep mode with &man.acpiconf.8;, the <option>-s</option>
2242 flag, and a <literal>1-5</literal> option. Most users will only need
2243 <literal>1</literal>. Option <literal>5</literal> will do a soft-off
2244 which is the same action as:</para>
2246 <screen>&prompt.root; <userinput>halt -p</userinput></screen>
2248 <para>The other options are available. Check out the &man.acpiconf.8;
2249 manual page for more information.</para>
2253 <sect1 id="ACPI-debug">
2257 <firstname>Nate</firstname>
2258 <surname>Lawson</surname>
2259 <contrib>Written by </contrib>
2264 <firstname>Peter</firstname>
2265 <surname>Schultz</surname>
2266 <contrib>With contributions from </contrib>
2269 <firstname>Tom</firstname>
2270 <surname>Rhodes</surname>
2275 <title>Using and Debugging &os; <acronym>ACPI</acronym></title>
2277 <para><acronym>ACPI</acronym> is a fundamentally new way of
2278 discovering devices, managing power usage, and providing
2279 standardized access to various hardware previously managed
2280 by the <acronym>BIOS</acronym>. Progress is being made toward
2281 <acronym>ACPI</acronym> working on all systems, but bugs in some
2282 motherboards' <firstterm><acronym>ACPI</acronym> Machine
2283 Language</firstterm> (<acronym>AML</acronym>) bytecode,
2284 incompleteness in &os;'s kernel subsystems, and bugs in the Intel
2285 <acronym>ACPI-CA</acronym> interpreter continue to appear.</para>
2287 <para>This document is intended to help you assist the &os;
2288 <acronym>ACPI</acronym> maintainers in identifying the root cause
2289 of problems you observe and debugging and developing a solution.
2290 Thanks for reading this and we hope we can solve your system's
2293 <sect2 id="ACPI-submitdebug">
2294 <title>Submitting Debugging Information</title>
2297 <para>Before submitting a problem, be sure you are running the latest
2298 <acronym>BIOS</acronym> version and, if available, embedded
2299 controller firmware version.</para>
2302 <para>For those of you that want to submit a problem right away,
2303 please send the following information to
2304 &a.bugs.name;</para>
2308 <para>Description of the buggy behavior, including system type
2309 and model and anything that causes the bug to appear. Also,
2310 please note as accurately as possible when the bug began
2311 occurring if it is new for you.</para>
2315 <para>The dmesg output after <quote>boot
2316 <option>-v</option></quote>, including any error messages
2317 generated by you exercising the bug.</para>
2321 <para>dmesg output from <quote>boot
2322 <option>-v</option></quote> with <acronym>ACPI</acronym>
2323 disabled, if disabling it helps fix the problem.</para>
2327 <para>Output from <quote>sysctl hw.acpi</quote>. This is also
2328 a good way of figuring out what features your system
2333 <para><acronym>URL</acronym> where your
2334 <firstterm><acronym>ACPI</acronym> Source Language</firstterm>
2335 (<acronym>ASL</acronym>)
2336 can be found. Do <emphasis>not</emphasis> send the
2337 <acronym>ASL</acronym> directly to the list as it can be
2338 very large. Generate a copy of your <acronym>ASL</acronym>
2339 by running this command:</para>
2341 <screen>&prompt.root; <userinput>acpidump -t -d > <replaceable>name</replaceable>-<replaceable>system</replaceable>.asl</userinput></screen>
2343 <para>(Substitute your login name for
2344 <replaceable>name</replaceable> and manufacturer/model for
2345 <replaceable>system</replaceable>. Example:
2346 <filename>njl-FooCo6000.asl</filename>)</para>
2352 <sect2 id="ACPI-background">
2353 <title>Background</title>
2355 <para><acronym>ACPI</acronym> is present in all modern computers
2356 that conform to the ia32 (x86), ia64 (Itanium), and amd64 (AMD)
2357 architectures. The full standard has many features including
2358 <acronym>CPU</acronym> performance management, power planes
2359 control, thermal zones, various battery systems, embedded
2360 controllers, and bus enumeration. Most systems implement less
2361 than the full standard. For instance, a desktop system usually
2362 only implements the bus enumeration parts while a laptop might
2363 have cooling and battery management support as well. Laptops
2364 also have suspend and resume, with their own associated
2367 <para>An <acronym>ACPI</acronym>-compliant system has various
2368 components. The <acronym>BIOS</acronym> and chipset vendors
2369 provide various fixed tables (e.g., <acronym>FADT</acronym>)
2370 in memory that specify things like the <acronym>APIC</acronym>
2371 map (used for <acronym>SMP</acronym>), config registers, and
2372 simple configuration values. Additionally, a table of bytecode
2373 (the <firstterm>Differentiated System Description Table</firstterm>
2374 <acronym>DSDT</acronym>) is provided that specifies a
2375 tree-like name space of devices and methods.</para>
2377 <para>The <acronym>ACPI</acronym> driver must parse the fixed
2378 tables, implement an interpreter for the bytecode, and modify
2379 device drivers and the kernel to accept information from the
2380 <acronym>ACPI</acronym> subsystem. For &os;, Intel has
2381 provided an interpreter (<acronym>ACPI-CA</acronym>) that is
2382 shared with Linux and NetBSD. The path to the
2383 <acronym>ACPI-CA</acronym> source code is
2384 <filename role="directory">src/sys/contrib/dev/acpica-unix-YYYYMMDD</filename>,
2385 where YYYYMMDD is the release date of the ACPI-CA source code. The
2386 glue code that allows <acronym>ACPI-CA</acronym> to work on
2387 &os; is in <filename>src/sys/dev/acpica5/Osd</filename>. Finally,
2388 drivers that implement various <acronym>ACPI</acronym> devices
2389 are found in <filename role="directory">src/sys/dev/acpica5</filename>,
2390 and architecture-dependent code resides in
2391 <filename role="directory">/sys/<replaceable>arch</replaceable>/acpica5</filename>.
2395 <sect2 id="ACPI-comprob">
2396 <title>Common Problems</title>
2398 <para>For <acronym>ACPI</acronym> to work correctly, all the parts
2399 have to work correctly. Here are some common problems, in order
2400 of frequency of appearance, and some possible workarounds or
2404 <title>Suspend/Resume</title>
2406 <para><acronym>ACPI</acronym> has three suspend to
2407 <acronym>RAM</acronym> (<acronym>STR</acronym>) states,
2408 <literal>S1</literal>-<literal>S3</literal>, and one suspend
2409 to disk state (<literal>STD</literal>), called
2410 <literal>S4</literal>. <literal>S5</literal> is
2411 <quote>soft off</quote> and is the normal state your system
2412 is in when plugged in but not powered up.
2413 <literal>S4</literal> can actually be implemented two separate
2414 ways. <literal>S4</literal><acronym>BIOS</acronym> is a
2415 <acronym>BIOS</acronym>-assisted suspend to disk.
2416 <literal>S4</literal><acronym>OS</acronym> is implemented
2417 entirely by the operating system.</para>
2419 <para>Start by checking <command>sysctl</command>
2420 <option>hw.acpi</option> for the suspend-related items. Here
2421 are the results for my Thinkpad:</para>
2423 <screen>hw.acpi.supported_sleep_state: S3 S4 S5</screen>
2424 <screen>hw.acpi.s4bios: 0</screen>
2426 <para>This means that I can use <command>acpiconf -s</command>
2427 to test <literal>S3</literal>,
2428 <literal>S4</literal><acronym>OS</acronym>, and
2429 <literal>S5</literal>. If <option>s4bios</option> was one
2430 (<literal>1</literal>), I would have
2431 <literal>S4</literal><acronym>BIOS</acronym>
2432 support instead of <literal>S4</literal>
2433 <acronym>OS</acronym>.</para>
2435 <para>When testing suspend/resume, start with
2436 <literal>S1</literal>, if supported. This state is most
2437 likely to work since it doesn't require much driver support.
2438 No one has implemented <literal>S2</literal> but if you have
2439 it, it's similar to <literal>S1</literal>. The next thing
2440 to try is <literal>S3</literal>. This is the deepest
2441 <acronym>STR</acronym> state and requires a lot of driver
2442 support to properly reinitialize your hardware. If you have
2443 problems resuming, feel free to email the &a.bugs.name; list but
2444 do not expect the problem to be resolved since there are a lot
2445 of drivers/hardware that need more testing and work.</para>
2447 <para>To help isolate the problem, remove as many drivers from
2448 your kernel as possible. If it works, you can narrow down
2449 which driver is the problem by loading drivers until it fails
2450 again. Typically binary drivers like
2451 <filename>nvidia.ko</filename>, <application>X11</application>
2452 display drivers, and <acronym>USB</acronym> will have the most
2453 problems while Ethernet interfaces usually work fine. If you
2454 can load/unload the drivers ok, you can automate this by
2455 putting the appropriate commands in
2456 <filename>/etc/rc.suspend</filename> and
2457 <filename>/etc/rc.resume</filename>. There is a
2458 commented-out example for unloading and loading a driver. Try
2459 setting <option>hw.acpi.reset_video</option> to zero (0) if
2460 your display is messed up after resume. Try setting longer or
2461 shorter values for <option>hw.acpi.sleep_delay</option> to see
2462 if that helps.</para>
2464 <para>Another thing to try is load a recent Linux distribution
2465 with <acronym>ACPI</acronym> support and test their
2466 suspend/resume support on the same hardware. If it works
2467 on Linux, it's likely a &os; driver problem and narrowing down
2468 which driver causes the problems will help us fix the problem.
2469 Note that the <acronym>ACPI</acronym> maintainers do not
2470 usually maintain other drivers (e.g sound,
2471 <acronym>ATA</acronym>, etc.) so any work done on tracking
2472 down a driver problem should probably eventually be posted
2473 to the &a.bugs.name; list and mailed to the driver
2474 maintainer. If you are feeling adventurous, go ahead and
2475 start putting some debugging &man.printf.3;s in a problematic
2476 driver to track down where in its resume function it
2479 <para>Finally, try disabling <acronym>ACPI</acronym> and
2480 enabling <acronym>APM</acronym> instead. If suspend/resume
2481 works with <acronym>APM</acronym>, you may be better off
2482 sticking with <acronym>APM</acronym>, especially on older
2483 hardware (pre-2000). It took vendors a while to get
2484 <acronym>ACPI</acronym> support correct and older hardware is
2485 more likely to have <acronym>BIOS</acronym> problems with
2486 <acronym>ACPI</acronym>.</para>
2490 <title>System Hangs (temporary or permanent)</title>
2492 <para>Most system hangs are a result of lost interrupts or an
2493 interrupt storm. Chipsets have a lot of problems based on how
2494 the <acronym>BIOS</acronym> configures interrupts before boot,
2495 correctness of the <acronym>APIC</acronym>
2496 (<acronym>MADT</acronym>) table, and routing of the
2497 <firstterm>System Control Interrupt</firstterm>
2498 (<acronym>SCI</acronym>).</para>
2500 <para>Interrupt storms can be distinguished from lost interrupts
2501 by checking the output of <command>vmstat -i</command>
2502 and looking at the line that has
2503 <literal>acpi0</literal>. If the counter is increasing at more
2504 than a couple per second, you have an interrupt storm. If the
2505 system appears hung, try breaking to <acronym>DDB</acronym>
2506 (<keycombo action="simul"><keycap>CTRL</keycap>
2507 <keycap>ALT</keycap><keycap>ESC</keycap></keycombo> on
2508 console) and type <option>show interrupts</option>.</para>
2510 <para>Your best hope when dealing with interrupt problems is to
2511 try disabling <acronym>APIC</acronym> support with
2512 <literal>hint.apic.0.disabled="1"</literal> in
2513 <filename>loader.conf</filename>.</para>
2517 <title>Panics</title>
2519 <para>Panics are relatively rare for <acronym>ACPI</acronym> and
2520 are the top priority to be fixed. The first step is to
2521 isolate the steps to reproduce the panic (if possible)
2522 and get a backtrace. Follow the advice for enabling
2523 <option>options DDB</option> and setting up a serial console
2524 (see <xref linkend="serialconsole-ddb">)
2525 or setting up a &man.dump.8; partition. You can get a
2526 backtrace in <acronym>DDB</acronym> with
2527 <option>tr</option>. If you have to handwrite the
2528 backtrace, be sure to at least get the lowest five (5) and top
2529 five (5) lines in the trace.</para>
2531 <para>Then, try to isolate the problem by booting with
2532 <acronym>ACPI</acronym> disabled. If that works, you can
2533 isolate the <acronym>ACPI</acronym> subsystem by using various
2534 values of <option>debug.acpi.disable</option>. See the
2535 &man.acpi.4; manual page for some examples.</para>
2539 <title>System Powers Up After Suspend or Shutdown</title>
2540 <para>First, try setting
2541 <option>hw.acpi.disable_on_poweroff=</option><quote>0</quote>
2542 in &man.loader.conf.5;. This keeps <acronym>ACPI</acronym>
2543 from disabling various events during the shutdown process.
2544 Some systems need this value set to <quote>1</quote> (the
2545 default) for the same reason. This usually fixes
2546 the problem of a system powering up spontaneously after a
2547 suspend or poweroff.</para>
2551 <title>Other Problems</title>
2553 <para>If you have other problems with <acronym>ACPI</acronym>
2554 (working with a docking station, devices not detected, etc.),
2555 please email a description to the mailing list as well;
2556 however, some of these issues may be related to unfinished
2557 parts of the <acronym>ACPI</acronym> subsystem so they might
2558 take a while to be implemented. Please be patient and
2559 prepared to test patches we may send you.</para>
2563 <sect2 id="ACPI-aslanddump">
2564 <title><acronym>ASL</acronym>, <command>acpidump</command>, and
2565 <acronym>IASL</acronym></title>
2567 <para>The most common problem is the <acronym>BIOS</acronym>
2568 vendors providing incorrect (or outright buggy!) bytecode. This
2569 is usually manifested by kernel console messages like
2572 <screen>ACPI-1287: *** Error: Method execution failed [\\_SB_.PCI0.LPC0.FIGD._STA] (Node 0xc3f6d160), AE_NOT_FOUND</screen>
2574 <para>Often, you can resolve these problems by updating your
2575 <acronym>BIOS</acronym> to the latest revision. Most console
2576 messages are harmless but if you have other problems like
2577 battery status not working, they're a good place to start
2578 looking for problems in the <acronym>AML</acronym>. The
2579 bytecode, known as <acronym>AML</acronym>, is compiled from a
2580 source language called <acronym>ASL</acronym>. The
2581 <acronym>AML</acronym> is found in the table known as the
2582 <acronym>DSDT</acronym>. To get a copy of your
2583 <acronym>ASL</acronym>, use &man.acpidump.8;. You should use
2584 both the <option>-t</option> (show contents of the fixed tables)
2585 and <option>-d</option> (disassemble <acronym>AML</acronym> to
2586 <acronym>ASL</acronym>) options. See the
2587 <link linkend="ACPI-submitdebug">Submitting Debugging
2588 Information</link> section for an example syntax.</para>
2590 <para>The simplest first check you can do is to recompile your
2591 <acronym>ASL</acronym> to check for errors. Warnings can
2592 usually be ignored but errors are bugs that will usually prevent
2593 <acronym>ACPI</acronym> from working correctly. To recompile
2594 your <acronym>ASL</acronym>, issue the following command:</para>
2596 <screen>&prompt.root; <userinput>iasl your.asl</userinput></screen>
2599 <sect2 id="ACPI-fixasl">
2600 <title>Fixing Your <acronym>ASL</acronym></title>
2602 <para>In the long run, our goal is for almost everyone to have
2603 <acronym>ACPI</acronym> work without any user intervention. At
2604 this point, however, we are still developing workarounds for
2605 common mistakes made by the <acronym>BIOS</acronym> vendors.
2606 The Microsoft interpreter (<filename>acpi.sys</filename> and
2607 <filename>acpiec.sys</filename>) does not strictly check for
2608 adherence to the standard, and thus many <acronym>BIOS</acronym>
2609 vendors who only test <acronym>ACPI</acronym> under Windows
2610 never fix their <acronym>ASL</acronym>. We hope to continue to
2611 identify and document exactly what non-standard behavior is
2612 allowed by Microsoft's interpreter and replicate it so &os; can
2613 work without forcing users to fix the <acronym>ASL</acronym>.
2614 As a workaround and to help us identify behavior, you can fix
2615 the <acronym>ASL</acronym> manually. If this works for you,
2616 please send a &man.diff.1; of the old and new
2617 <acronym>ASL</acronym> so we can possibly work around the buggy
2618 behavior in <acronym>ACPI-CA</acronym> and thus make your fix
2621 <para>Here is a list of common error messages, their cause, and
2622 how to fix them:</para>
2625 <title>_OS dependencies</title>
2627 <para>Some <acronym>AML</acronym> assumes the world consists of
2628 various Windows versions. You can tell &os; to claim it is
2629 any <acronym>OS</acronym> to see if this fixes problems you
2630 may have. An easy way to override this is to set
2631 <option>hw.acpi.osname</option>=<quote>Windows 2001</quote>
2632 in <filename>/boot/loader.conf</filename> or other similar
2633 strings you find in the <acronym>ASL</acronym>.</para>
2636 <title>Missing Return statements</title>
2638 <para>Some methods do not explicitly return a value as the
2639 standard requires. While <acronym>ACPI-CA</acronym>
2640 does not handle this, &os; has a workaround that allows it to
2641 return the value implicitly. You can also add explicit
2642 Return statements where required if you know what value should
2643 be returned. To force <command>iasl</command> to compile the
2644 <acronym>ASL</acronym>, use the <option>-f</option>
2649 <title>Overriding the Default <acronym>AML</acronym></title>
2651 <para>After you customize <filename>your.asl</filename>, you
2652 will want to compile it, run:</para>
2654 <screen>&prompt.root; <userinput>iasl your.asl</userinput></screen>
2656 <para>You can add the <option>-f</option> flag to force creation
2657 of the <acronym>AML</acronym>, even if there are errors during
2658 compilation. Remember that some errors (e.g., missing Return
2659 statements) are automatically worked around by the
2662 <para><filename>DSDT.aml</filename> is the default output
2663 filename for <command>iasl</command>. You can load this
2664 instead of your <acronym>BIOS</acronym>'s buggy copy (which
2665 is still present in flash memory) by editing
2666 <filename>/boot/loader.conf</filename> as
2669 <programlisting>acpi_dsdt_load="YES"
2670 acpi_dsdt_name="/boot/DSDT.aml"</programlisting>
2672 <para>Be sure to copy your <filename>DSDT.aml</filename> to the
2673 <filename role="directory">/boot</filename> directory.</para>
2676 <sect2 id="ACPI-debugoutput">
2677 <title>Getting Debugging Output From
2678 <acronym>ACPI</acronym></title>
2680 <para>The <acronym>ACPI</acronym> driver has a very flexible
2681 debugging facility. It allows you to specify a set of subsystems
2682 as well as the level of verbosity. The subsystems you wish to
2683 debug are specified as <quote>layers</quote> and are broken down
2684 into <acronym>ACPI-CA</acronym> components (ACPI_ALL_COMPONENTS)
2685 and <acronym>ACPI</acronym> hardware support (ACPI_ALL_DRIVERS).
2686 The verbosity of debugging output is specified as the
2687 <quote>level</quote> and ranges from ACPI_LV_ERROR (just report
2688 errors) to ACPI_LV_VERBOSE (everything). The
2689 <quote>level</quote> is a bitmask so multiple options can be set
2690 at once, separated by spaces. In practice, you will want to use
2691 a serial console to log the output if it is so long
2692 it flushes the console message buffer. </para>
2694 <para>Debugging output is not enabled by default. To enable it,
2695 add <option>options ACPI_DEBUG</option> to your kernel config
2696 if <acronym>ACPI</acronym> is compiled into the kernel. You can
2697 add <option>ACPI_DEBUG=1</option> to your
2698 <filename>/etc/make.conf</filename> to enable it globally. If
2699 it is a module, you can recompile just your
2700 <filename>acpi.ko</filename> module as follows:</para>
2702 <screen>&prompt.root; <userinput>cd /sys/dev/acpica5
2703 && make clean &&
2704 make ACPI_DEBUG=1</userinput></screen>
2706 <para>Install <filename>acpi.ko</filename> in
2707 <filename role="directory">/boot/kernel</filename> and add your
2708 desired level and layer to <filename>loader.conf</filename>.
2709 This example enables debug messages for all
2710 <acronym>ACPI-CA</acronym> components and all
2711 <acronym>ACPI</acronym> hardware drivers
2712 (<acronym>CPU</acronym>, <acronym>LID</acronym>, etc.) It will
2713 only output error messages, the least verbose level.</para>
2715 <programlisting>debug.acpi.layer="ACPI_ALL_COMPONENTS ACPI_ALL_DRIVERS"
2716 debug.acpi.level="ACPI_LV_ERROR"</programlisting>
2718 <para>If the information you want is triggered by a specific event
2719 (say, a suspend and then resume), you can leave out changes to
2720 <filename>loader.conf</filename> and instead use
2721 <command>sysctl</command> to specify the layer and level after
2722 booting and preparing your system for the specific event. The
2723 <command>sysctl</command>s are named the same as the tunables
2724 in <filename>loader.conf</filename>.</para>
2727 <sect2 id="ACPI-References">
2728 <title>References</title>
2730 <para>More information about <acronym>ACPI</acronym> may be found
2731 in the following locations:</para>
2735 <para>The FreeBSD &a.acpi; (This is FreeBSD-specific; posting
2736 &os; questions here may not generate much of an answer.)</para>
2740 <para>The <acronym>ACPI</acronym> Mailing List Archives (FreeBSD)
2741 <ulink url="http://lists.freebsd.org/pipermail/freebsd-acpi/"></ulink></para>
2745 <para>The old <acronym>ACPI</acronym> Mailing List Archives (FreeBSD)
2746 <ulink url="http://home.jp.FreeBSD.org/mail-list/acpi-jp/"></ulink></para>
2750 <para>The <acronym>ACPI</acronym> 2.0 Specification
2751 <ulink url="http://acpi.info/spec.htm"></ulink></para>
2755 <para>&os; Manual pages:
2758 &man.acpidb.8;</para>
2763 url="http://www.cpqlinux.com/acpi-howto.html#fix_broken_dsdt">
2764 <acronym>DSDT</acronym> debugging resource</ulink>.
2765 (Uses Compaq as an example but generally useful.)</para>
2775 sgml-declaration: "../chapter.decl"
2778 sgml-always-quote-attributes: t
2779 sgml-parent-document: ("../book.sgml" "part" "chapter")