path: root/en_US.ISO8859-1/books/handbook/mac/chapter.xml

<?xml version="1.0" encoding="iso-8859-1" standalone="no"?>
<!--
     The FreeBSD Documentation Project
     $FreeBSD$
-->

<chapter id="mac">
  <chapterinfo>
    <authorgroup>
      <author>
	<firstname>Tom</firstname>
	<surname>Rhodes</surname>
	<contrib>Written by </contrib>
      </author>
    </authorgroup>
  </chapterinfo>

  <title>Mandatory Access Control</title>

  <sect1 id="mac-synopsis">
    <title>Synopsis</title>

    <indexterm><primary>MAC</primary></indexterm>
    <indexterm>
      <primary>Mandatory Access Control</primary>
      <see>MAC</see>
    </indexterm>

    <para>&os;&nbsp;5.X introduced new security extensions from the
      <ulink url="http://www.trustedbsd.org">TrustedBSD
	Project</ulink> based on the &posix;.1e draft.  Two of the
      most significant new security mechanisms are file system Access
      Control Lists (<acronym>ACL</acronym>s) and Mandatory Access
      Control (<acronym>MAC</acronym>) facilities.  MAC allows new
      access control modules to be loaded, implementing new security
      policies.  Some modules provide protections for a narrow subset
      of the system, hardening a particular service.  Others provide
      comprehensive labeled security across all subjects and objects.
      The mandatory part of the definition indicates that enforcement
      of controls is performed by administrators and the operating
      system.  This is in contrast to the default security mechanism
      of Discretionary Access Control (<acronym>DAC</acronym> where
      enforcement is left to the discretion of users.</para>

    <para>This chapter focuses on the <acronym>MAC</acronym> framework
      and the set of pluggable security policy modules &os; provides
      for enabling various security mechanisms.</para>

    <para>After reading this chapter, you will know:</para>

    <itemizedlist>
      <listitem>
	<para>Which <acronym>MAC</acronym> security policy modules
	  are included in &os; and their associated mechanisms.</para>
      </listitem>

      <listitem>
	<para>The capabilities of <acronym>MAC</acronym> security
	  policy modules as well as the difference between a labeled
	  and non-labeled policy.</para>
      </listitem>

      <listitem>
	<para>How to efficiently configure a system to use the
	  <acronym>MAC</acronym> framework.</para>
      </listitem>

      <listitem>
	<para>How to configure the different security policy modules
	  included with the <acronym>MAC</acronym> framework.</para>
      </listitem>

      <listitem>
	<para>How to implement a more secure environment using the
	  <acronym>MAC</acronym> framework.</para>
      </listitem>

      <listitem>
	<para>How to test the <acronym>MAC</acronym> configuration
	  to ensure the framework has been properly
	  implemented.</para>
      </listitem>
    </itemizedlist>

    <para>Before reading this chapter, you should:</para>

    <itemizedlist>
      <listitem>
	<para>Understand &unix; and &os; basics
	  (<xref linkend="basics"/>).</para>
      </listitem>

      <listitem>
	<para>Have some familiarity with security and how it
	  pertains to &os; (<xref linkend="security"/>).</para>
      </listitem>
    </itemizedlist>

    <warning>
      <para>Improper <acronym>MAC</acronym> configuration may cause
	loss of system access, aggravation of users, or inability to
	access the features provided by
	<application>Xorg</application>.  More importantly,
	<acronym>MAC</acronym> should not be relied upon to completely
	secure a system.  The <acronym>MAC</acronym> framework only
	augments an existing security policy.  Without sound security
	practices and regular security checks, the system will never
	be completely secure.</para>

      <para>The examples contained within this chapter are for
	demonstration purposes and the example settings should
	<emphasis>not</emphasis> be implemented on a production
	system.  Implementing any security policy takes a good deal of
	understanding, proper design, and thorough testing.</para>
    </warning>

    <sect2>
      <title>What Will Not Be Covered</title>

      <para>This chapter covers a broad range of security issues
	relating to the <acronym>MAC</acronym> framework.  The
	development of new <acronym>MAC</acronym> security policy
	modules will not be covered.  A number of security policy
	modules included with the <acronym>MAC</acronym> framework
	have specific characteristics which are provided for both
	testing and new module development.  These include
	&man.mac.test.4;, &man.mac.stub.4; and &man.mac.none.4;.
	For more information on these security policy modules and
	the various mechanisms they provide, refer to their manual
	pages.</para>
    </sect2>
  </sect1>

  <sect1 id="mac-inline-glossary">
    <title>Key Terms in This Chapter</title>

    <para>Before reading this chapter, a few key terms must be
      explained:</para>

    <itemizedlist>
      <listitem>
	<para><emphasis>compartment</emphasis>: a set of programs and
	  data to be partitioned or separated, where users are given
	  explicit access to specific component of a system.  A
	  compartment represents a grouping, such as a work group,
	  department, project, or topic.  Compartments make it
	  possible to implement a need-to-know-basis security
	  policy.</para>
      </listitem>

      <listitem>
	<para><emphasis>high-watermark</emphasis>: this type of
	  policy permits the raising of security levels for the
	  purpose of accessing higher level information.  In most
	  cases, the original level is restored after the process
	  is complete.  Currently, the &os; <acronym>MAC</acronym>
	  framework does not include this type of policy.</para>
      </listitem>

      <listitem>
	<para><emphasis>integrity</emphasis>: the level of trust which
	  can be placed on data.  As the integrity of the data is
	  elevated, so does the ability to trust that data.</para>
      </listitem>

      <listitem>
	<para><emphasis>label</emphasis>: a security attribute which
	  can be applied to files, directories, or other items in the
	  system.  It could be considered a confidentiality stamp.
	  When a label is placed on a file, it describes the security
	  properties of that file and will only permit access by
	  files, users, and resources with a similar security setting.
	  The meaning and interpretation of label values depends on
	  the policy configuration.  Some policies treat a label as
	  representing the integrity or secrecy of an object while
	  other policies might use labels to hold rules for
	  access.</para>
      </listitem>

      <listitem>
	<para><emphasis>level</emphasis>: the increased or decreased
	  setting of a security attribute.  As the level increases,
	  its security is considered to elevate as well.</para>
      </listitem>

      <listitem>
	<para><emphasis>low-watermark</emphasis>: this type of
	  policy permits lowering security levels for the purpose of
	  accessing information which is less secure.  In most cases,
	  the original security level of the user is restored after
	  the process is complete.  The only security policy module in
	  &os; to use this is &man.mac.lomac.4;.</para>
      </listitem>

      <listitem>
	<para><emphasis>multilabel</emphasis>: this property is a file
	  system option which can be set in single user mode using
	  &man.tunefs.8;, during boot using &man.fstab.5;, or during
	  the creation of a new file system.  This option permits
	  an administrator to apply different <acronym>MAC</acronym>
	  labels on different objects.  This option only applies to
	  security policy modules which support labeling.</para>
      </listitem>

      <listitem>
	<para><emphasis>object</emphasis>: an entity through which
	  information flows under the direction of a
	  <emphasis>subject</emphasis>.  This includes directories,
	  files, fields, screens, keyboards, memory, magnetic storage,
	  printers or any other data storage or moving device.  An
	  object is a data container or a system resource.  Access to
	  an <emphasis>object</emphasis> effectively means access to
	  its data.</para>
      </listitem>

      <listitem>
	<para><emphasis>policy</emphasis>: a collection of rules
	  which defines how objectives are to be achieved.  A
	  <emphasis>policy</emphasis> usually documents how certain
	  items are to be handled.  This chapter considers the term
	  <emphasis>policy</emphasis> to be a <emphasis>security
	  policy</emphasis>, or a collection of rules which controls
	  the flow of data and information and defines who has access
	  to that data and information.</para>
      </listitem>

      <listitem>
	<para><emphasis>sensitivity</emphasis>: usually used when
	  discussing Multilevel Security <acronym>MLS</acronym>.  A
	  sensitivity level describes how important or secret the data
	  should be.  As the sensitivity level increases, so does the
	  importance of the secrecy, or confidentiality, of the
	  data.</para>
      </listitem>

      <listitem>
	<para><emphasis>single label</emphasis>: a policy where the
	  entire file system uses one label to enforce access control
	  over the flow of data.  Whenever <option>multilabel</option>
	  is not set, all files will conform to the same label
	  setting.</para>
      </listitem>

      <listitem>
	<para><emphasis>subject</emphasis>: any active entity that
	  causes information to flow between
	  <emphasis>objects</emphasis> such as a user, user process,
	  or system process.  On &os;, this is almost always a
	  thread acting in a process on behalf of a user.</para>
      </listitem>
    </itemizedlist>
  </sect1>

  <sect1 id="mac-initial">
    <title>Explanation of MAC</title>

    <para>With all of these new terms in mind, consider how the
      <acronym>MAC</acronym> framework augments the security of
      the system as a whole.  The various security policy modules
      provided by the <acronym>MAC</acronym> framework could be used
      to protect the network and file systems or to block users from
      accessing certain ports and sockets.  Perhaps the best use of
      the policy modules is to load several security policy modules at
      a time in order to provide a <acronym>MLS</acronym> environment.
      This approach differs from a hardening policy, which typically
      hardens elements of a system which are used only for specific
      purposes.  The downside to <acronym>MLS</acronym> is increased
      administrative overhead.</para>

    <para>The overhead is minimal when compared to the lasting effect
      of a framework which provides the ability to pick and choose
      which policies are required for a specific configuration and
      which keeps performance overhead down.  The reduction of support
      for unneeded policies can increase the overall performance of
      the system as well as offer flexibility of choice.  A good
      implementation would consider the overall security requirements
      and effectively implement the various security policy modules
      offered by the framework.</para>

    <para>A system utilizing <acronym>MAC</acronym> guarantees that a
      user will not be permitted to change security attributes at
      will.  All user utilities, programs, and scripts must work
      within the constraints of the access rules provided by the
      selected security policy modules and total control of the
      <acronym>MAC</acronym> access rules are in the hands of the
      system administrator.</para>

    <para>It is the duty of the system administrator to
      carefully select the correct security policy modules.  For an
      environment that needs to limit access control over the network,
      the &man.mac.portacl.4;, &man.mac.ifoff.4;, and &man.mac.biba.4;
      policy modules make good starting points.  For an environment
      where strict confidentiality of file system objects is required,
      consider the &man.mac.bsdextended.4; and &man.mac.mls.4; policy
      modules.</para>

    <para>Policy decisions could be made based on network
      configuration.  If only certain users should be permitted
      access to &man.ssh.1;, the &man.mac.portacl.4; policy module is
      a good choice.  In the case of file systems, access to objects
      might be considered confidential to some users, but not to
      others.  As an example, a large development team might be
      broken off into smaller projects where  developers in project A
      might not be permitted to access objects written by developers
      in project B.  Yet both projects might need to access objects
      created by developers in project C.  Using the different
      security policy modules provided by the <acronym>MAC</acronym>
      framework, users could be divided into these groups and then
      given access to the appropriate objects.</para>

    <para>Each security policy module has a unique way of dealing with
      the overall security of a system.  Module selection should be
      based on a well thought out security policy which may require
      revision and reimplementation.  Understanding the different
      security policy modules offered by the <acronym>MAC</acronym>
      framework will help administrators choose the best policies
      for their situations.</para>

    <caution>
      <para>Implementing <acronym>MAC</acronym> is much like
	implementing a firewall, care must be taken to prevent being
	completely locked out of the system.  The ability to revert
	back to a previous configuration should be considered and the
	implementation of <acronym>MAC</acronym> remotely should be
	done with extreme caution.</para>
    </caution>
  </sect1>

  <sect1 id="mac-understandlabel">
    <title>Understanding MAC Labels</title>

    <para>A <acronym>MAC</acronym> label is a security attribute
      which may be applied to subjects and objects throughout
      the system.</para>

    <para>When setting a label, the administrator must be able to
      comprehend what exactly is being done and understand any
      implications in order to prevent unexpected or undesired
      behavior of the system.  The attributes available on an object
      depend on the loaded policy module as policy modules interpret
      their attributes in different ways.</para>

    <para>The security label on an object is used as a part of a
      security access control decision by a policy.  With some
      policies, the label contains all of the information necessary
      to make a decision.  In other policies, the labels may be
      processed as part of a larger rule set.  For instance, setting
      the label of <literal>biba/low</literal> on a file will
      represent a label maintained by the Biba security policy module,
      with a value of <quote>low</quote>.</para>

    <para>A few policy modules which support the labeling feature
      in &os; offer three specific predefined labels:  low, high, and
      equal.  Such policy modules enforce access control in a
      different manner with each policy module, where the low label is
      the lowest setting, the equal label sets the subject or object
      to be disabled or unaffected, and the high label enforces the
      highest setting available in the Biba and <acronym>MLS</acronym>
      policy modules.</para>

    <para>Within single label file system environments, only one
      label may be used on objects.  This label enforces one set of
      access permissions across the entire system and in many
      environments may be all that is required.  There are a few
      cases where multiple labels may be set on objects or subjects
      in the file system by passing <option>multilabel</option> to
      &man.tunefs.8;.</para>

    <para>In the case of Biba and <acronym>MLS</acronym>, a numeric
      label may be set to indicate the precise level of hierarchical
      control.  This numeric level is used to partition or sort
      information into different groups of classification only
      permitting access to that group or a higher group level.</para>

    <para>In most cases, the administrator will set up a single label
      to use throughout the file system.  This is similar to
      <acronym>DAC</acronym> to some extent as
      <username>root</username> is the one in control and who
      configures the policies so that users are placed in the
      appropriate categories/access levels.  Alas, many policy modules
      can restrict the <username>root</username> user as well.  Basic
      control over objects will then be released to the group, but
      <username>root</username> may revoke or modify the settings
      at any time.  This is the hierarchical/clearance model covered
      by policies such as Biba and <acronym>MLS</acronym>.</para>

    <sect2>
      <title>Label Configuration</title>

      <para>Virtually all aspects of label policy module configuration
	will be performed using the base system utilities.  These
	commands provide a simple interface for object or subject
	configuration or the manipulation and verification of
	the configuration.</para>

      <para>All configuration may be done using &man.setfmac.8; and
	&man.setpmac.8;.  <command>setfmac</command> is used to set
	<acronym>MAC</acronym> labels on system objects while
	<command>setpmac</command> is used to set the labels on system
	subjects.  Observe:</para>

      <screen>&prompt.root; <userinput>setfmac biba/high test</userinput></screen>

      <para>If the configuration is successful, the prompt will be
	returned without error.  A common error is
	<errorname>Permission denied</errorname> which usually occurs
	when the label is being set or modified on an object which is
	restricted.<footnote>Other conditions may produce different
	  failures.  For instance, the file may not be owned by the
	  user attempting to relabel the object, the object may not
	  exist, or the object may be read only.  A mandatory policy
	  will not allow the process to relabel the file, maybe
	  because of a property of the file, a property of the
	  process, or a property of the proposed new label value.  For
	  example, a user running at low integrity tries to change the
	  label of a high integrity file.  Or perhaps a user running
	  at low integrity tries to change the label of a low
	  integrity file to a high integrity label.</footnote>  The
	system administrator may use the following commands to
	overcome this:</para>

      <screen>&prompt.root; <userinput>setfmac biba/high test</userinput>
<errorname>Permission denied</errorname>
&prompt.root; <userinput>setpmac biba/low setfmac biba/high test</userinput>
&prompt.root; <userinput>getfmac test</userinput>
test: biba/high</screen>

      <para><command>setpmac</command> can be used to override the
	policy module's settings by assigning a different label to the
	invoked process.  <command>getpmac</command> is usually used
	with currently running processes, such as
	<application>sendmail</application>.  It takes a process ID in
	place of a command.  If users attempt to manipulate a file not
	in their access, subject to the rules of the loaded policy
	modules, the <errorname>Operation not permitted</errorname>
	error will be displayed by the
	<function>mac_set_link</function> function.</para>

      <sect3>
	<title>Common Label Types</title>

	<para>For the &man.mac.biba.4;, &man.mac.mls.4; and
	  &man.mac.lomac.4; policy modules, the ability to assign
	  simple labels is provided.  These take the form of high,
	  equal, and low, where:</para>

	<itemizedlist>
	  <listitem>
	    <para>The <literal>low</literal> label is considered the
	      lowest label setting an object or subject may have.
	      Setting this on objects or subjects blocks their access
	      to objects or subjects marked high.</para>
	  </listitem>

	  <listitem>
	    <para>The <literal>equal</literal> label should only be
	      placed on objects considered to be exempt from the
	      policy.</para>
	  </listitem>

	  <listitem>
	    <para>The <literal>high</literal> label grants an object
	      or subject the highest possible setting.</para>
	  </listitem>
	</itemizedlist>

	<para>With respect to each policy module, each of those
	  settings will establish a different information flow
	  directive.  Refer to the manual pages of the module to
	  determine the traits of these generic label
	  configurations.</para>

	<sect4>
	  <title>Advanced Label Configuration</title>

	  <para>Numeric grade labels are used for
	    <literal>comparison:compartment+compartment</literal>.
	    For example:</para>

	  <programlisting>biba/10:2+3+6(5:2+3-20:2+3+4+5+6)</programlisting>

	  <para>may be interpreted as <quote>Biba Policy
	      Label</quote>/<quote>Grade
	      10</quote>:<quote>Compartments 2, 3 and 6</quote>:
	    (<quote>grade 5 ...</quote>)</para>

	  <para>In this example, the first grade would be considered
	    the <quote>effective grade</quote> with
	    <quote>effective compartments</quote>, the second grade
	    is the low grade, and the last one is the high grade.
	    In most configurations, these settings will not be used
	    as they are advanced configurations.</para>

	  <para>System objects only have a current grade/compartment.
	    System subjects reflect the range of available rights in
	    the system, and network interfaces, where they are used
	    for access control.</para>

	  <para>The grade and compartments in a subject and object
	    pair are used to construct a relationship known as
	    <quote>dominance</quote>, in which a subject dominates an
	    object, the object dominates the subject, neither
	    dominates the other, or both dominate each other.  The
	    <quote>both dominate</quote> case occurs when the two
	    labels are equal.  Due to the information flow nature of
	    Biba, a user has rights to a set of compartments that
	    might correspond to projects, but objects also have a set
	    of compartments.  Users may have to subset their rights
	    using <command>su</command> or <command>setpmac</command>
	    in order to access objects in a compartment from which
	    they are not restricted.</para>
	</sect4>
      </sect3>

      <sect3>
	<title>Users and Label Settings</title>

	<para>Users are required to have labels so that their files
	  and processes properly interact with the security policy
	  defined on the system.  This is configured in
	  <filename>login.conf</filename> using login classes.  Every
	  policy module that uses labels will implement the user class
	  setting.</para>

	<para>An example entry containing every policy module setting
	  is displayed below:</para>

	<programlisting>default:\
	:copyright=/etc/COPYRIGHT:\
	:welcome=/etc/motd:\
	:setenv=MAIL=/var/mail/$,BLOCKSIZE=K:\
	:path=~/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin:\
	:manpath=/usr/share/man /usr/local/man:\
	:nologin=/usr/sbin/nologin:\
	:cputime=1h30m:\
	:datasize=8M:\
	:vmemoryuse=100M:\
	:stacksize=2M:\
	:memorylocked=4M:\
	:memoryuse=8M:\
	:filesize=8M:\
	:coredumpsize=8M:\
	:openfiles=24:\
	:maxproc=32:\
	:priority=0:\
	:requirehome:\
	:passwordtime=91d:\
	:umask=022:\
	:ignoretime@:\
	:label=partition/13,mls/5,biba/10(5-15),lomac/10[2]:</programlisting>

	<para>To set the
	  user class default label which will be enforced by
	  <acronym>MAC</acronym>, use <option>label</option>.  Users
	  are never permitted to modify this value.  In a real
	  configuration, however, the administrator would never enable
	  every policy module.  It is recommended that the rest of
	  this chapter be reviewed before any configuration is
	  implemented.</para>

	<note>
	  <para>Users may change their label after they login, subject
	    to the constraints of the policy.  The example above tells
	    the Biba policy that a process's minimum integrity is 5,
	    its maximum is 15, and the default effective label is 10.
	    The process will run at 10 until it chooses to change
	    label, perhaps due to the user using &man.setpmac.8;,
	    which will be constrained by Biba to the configured
	    range.</para>
	</note>

	<para>After any change to
	  <filename>login.conf</filename>, the login class capability
	  database must be rebuilt using
	  <command>cap_mkdb</command>.</para>

	<para>Many sites have a large number of users requiring
	  several different user classes.  In depth planning is
	  required as this may get extremely difficult to
	  manage.</para>
      </sect3>

      <sect3>
	<title>Network Interfaces and Label Settings</title>

	<para>Labels may be set on network interfaces to help
	  control the flow of data across the network.  Policies
	  using network interface labels function in the same way that
	  policies function with respect to objects.  Users at high
	  settings in <literal>biba</literal>, for example, will not
	  be permitted to access network interfaces with a label of
	  low.</para>

	<para><option>maclabel</option> may be passed to
	  <command>ifconfig</command> when setting the
	  <acronym>MAC</acronym> label on network interfaces.  For
	  example:</para>

	<screen>&prompt.root; <userinput>ifconfig bge0 maclabel biba/equal</userinput></screen>

	<para>will set the <acronym>MAC</acronym> label of
	  <literal>biba/equal</literal> on the &man.bge.4; interface.
	  When using a setting similar to
	  <literal>biba/high(low-high)</literal>, the entire label
	  should be quoted to prevent an error from being
	  returned.</para>

	<para>Each policy module which supports labeling has a tunable
	  which may be used to disable the <acronym>MAC</acronym>
	  label on network interfaces.  Setting the label to
	  <option>equal</option> will have a similar effect.  Review
	  the output of <command>sysctl</command>, the policy manual
	  pages, and the information in this chapter for more
	  information on those tunables.</para>
      </sect3>
    </sect2>

    <sect2>
      <title>Singlelabel or Multilabel?</title>

      <para>By default, the system will use
	<option>singlelabel</option>.  For the administrator, there
	are several differences which offer pros and cons to the
	flexibility in the system's security model.</para>

      <para>A security policy which uses <option>singlelabel</option>
	only permits one label, such as <literal>biba/high</literal>,
	to be used for each subject or object.  This provides lower
	administration overhead, but decreases the flexibility of
	policies which support labeling.</para>

      <para><option>multilabel</option> permits each subject or object
	to have its own independent <acronym>MAC</acronym> label.
	The decision to use <option>multilabel</option> or
	<option>singlelabel</option> is only required for the policies
	which implement the labeling feature, including the Biba,
	Lomac, and <acronym>MLS</acronym> policies.</para>

      <para>In many cases, <option>multilabel</option> may not be
	needed.  Consider the following situation and security
	model:</para>

      <itemizedlist>
	<listitem>
	  <para>&os; web-server using the <acronym>MAC</acronym>
	    framework and a mix of the various policies.</para>
	</listitem>

	<listitem>
	  <para>This machine only requires one label,
	    <literal>biba/high</literal>, for everything in the
	    system.  This file system would not require
	    <option>multilabel</option> as a single label will always
	    be in effect.</para>
	</listitem>

	<listitem>
	  <para>But, this machine will be a web server and should
	    have the web server run at <literal>biba/low</literal>
	    to prevent write up capabilities.  The server could
	    use a separate partition set at
	    <literal>biba/low</literal> for most if not all
	    of its runtime state.</para>
	</listitem>
      </itemizedlist>

      <para>If any of the non-labeling policies are to be used,
	<option>multilabel</option> would not be required.  These
	include the <literal>seeotheruids</literal>,
	<literal>portacl</literal> and <literal>partition</literal>
	policies.</para>

      <para>Using <option>multilabel</option> with a partition and
	establishing a security model based on
	<option>multilabel</option> functionality could increase
	administrative overhead as everything in the file system has a
	label.  This includes directories, files, and even device
	nodes.</para>

      <para>The following command will set <option>multilabel</option>
	on the file systems to have multiple labels.  This may only be
	done in single user mode and is not a requirement for the swap
	file system:</para>

      <screen>&prompt.root; <userinput>tunefs -l enable /</userinput></screen>

      <note>
	<para>Some users have experienced problems with setting the
	  <option>multilabel</option> flag on the root partition.
	  If this is the case, please review the
	  <xref linkend="mac-troubleshoot"/> of this chapter.</para>
      </note>
    </sect2>
  </sect1>

  <sect1 id="mac-planning">
    <title>Planning the Security Configuration</title>

    <para>Whenever a new technology is implemented, a planning phase
      is recommended.  During the planning stages, an administrator
      should consider the implementation requirements and the
      implementation goals.</para>

    <para>For <acronym>MAC</acronym> installations, these
      include:</para>

    <itemizedlist>
      <listitem>
	<para>How to classify information and resources available on
	  the target systems.</para>
      </listitem>

      <listitem>
	<para>Which information or resources to restrict access to
	  along with the type of restrictions that should be
	  applied.</para>
      </listitem>

      <listitem>
	<para>Which <acronym>MAC</acronym> module or modules will be
	  required to achieve this goal.</para>
      </listitem>
    </itemizedlist>

    <para>Good planning helps to ensure a trouble-free and efficient
      trusted system implementation.  A trial run of the trusted
      system and its configuration should occur
      <emphasis>before</emphasis> a <acronym>MAC</acronym>
      implementation is used on production systems.  The idea of
      just letting loose on a system with <acronym>MAC</acronym> is
      like setting up for failure.</para>

    <para>Different environments have different needs and
      requirements.  Establishing an in depth and complete security
      profile will decrease the need of changes once the system
      goes live.  The rest of this chapter covers the available
      modules, describes their use and configuration, and in some
      cases, provides insight on applicable situations.  For instance,
      a web server might use the &man.mac.biba.4; and
      &man.mac.bsdextended.4; policies.  In the case of a machine
      with few local users, &man.mac.partition.4; might be a good
      choice.</para>
  </sect1>

  <sect1 id="mac-modules">
    <title>Module Configuration</title>

    <para>Beginning with &os;&nbsp;8.0, the default &os; kernel
      includes <literal>options MAC</literal>. This means that
      every module included with the <acronym>MAC</acronym>
      framework may be loaded as a run-time kernel module.  The
      recommended method is to add the module name to
      <filename>/boot/loader.conf</filename> so that it will load
      during boot.  Each module also provides a kernel option
      for those administrators who choose to compile their own
      custom kernel.</para>

    <para>Some modules support the use of labeling, which is
      controlling access by enforcing a label such as <quote>this is
      allowed and this is not</quote>.  A label configuration file may
      control how files may be accessed, network communication can be
      exchanged, and more.  The previous section showed how the
      <option>multilabel</option> flag could be set on file systems to
      enable per-file or per-partition access control.</para>

    <para>A single label configuration enforces only one label
      across the system, that is why the <command>tunefs</command>
      option is called <option>multilabel</option>.</para>
  </sect1>

  <sect1 id="mac-seeotheruids">
    <title>The &man.mac.seeotheruids.4; Module</title>

    <indexterm>
      <primary>MAC See Other UIDs Policy</primary>
    </indexterm>
    <para>Module name: <filename>mac_seeotheruids.ko</filename></para>

    <para>Kernel configuration line:
      <literal>options MAC_SEEOTHERUIDS</literal></para>

    <para>Boot option:
      <literal>mac_seeotheruids_load="YES"</literal></para>

    <para>The &man.mac.seeotheruids.4; module mimics and extends
      the <varname>security.bsd.see_other_uids</varname> and
      <varname>security.bsd.see_other_gids</varname>
      <command>sysctl</command> tunables.  This option does
      not require any labels to be set before configuration and
      can operate transparently with the other modules.</para>

    <para>After loading the module, the following
      <command>sysctl</command> tunables may be used to control
      the features:</para>

    <itemizedlist>
      <listitem>
	<para><varname>security.mac.seeotheruids.enabled</varname>
	  enables the module and uses the default settings which deny
	  users the ability to view processes and sockets owned by
	  other users.</para>
      </listitem>

      <listitem>
	<para>
	  <varname>security.mac.seeotheruids.specificgid_enabled</varname>
	  allows certain groups to be exempt from this policy.  To
	  exempt specific groups from this policy, use the
	  <varname>security.mac.seeotheruids.specificgid=<replaceable>XXX</replaceable></varname>
	  <command>sysctl</command> tunable.  Replace
	  <replaceable>XXX</replaceable> with the numeric group ID to
	  be exempted.</para>
      </listitem>

      <listitem>
	<para>
	  <varname>security.mac.seeotheruids.primarygroup_enabled</varname>
	  is used to exempt specific primary groups from this policy.
	  When using this tunable,
	  <varname>security.mac.seeotheruids.specificgid_enabled</varname>
	  may not be set.</para>
      </listitem>
    </itemizedlist>
  </sect1>

  <sect1 id="mac-bsdextended">
    <title>The &man.mac.bsdextended.4; Module</title>

    <indexterm>
      <primary>MAC</primary>
	<secondary>File System Firewall Policy</secondary>
    </indexterm>
    <para>Module name: <filename>mac_bsdextended.ko</filename></para>

    <para>Kernel configuration line:
      <literal>options MAC_BSDEXTENDED</literal></para>

    <para>Boot option:
      <literal>mac_bsdextended_load="YES"</literal></para>

    <para>The &man.mac.bsdextended.4; module enforces the file system
      firewall.  This module's policy provides an extension to the
      standard file system permissions model, permitting an
      administrator to create a firewall-like ruleset to protect
      files, utilities, and directories in the file system hierarchy.
      When access to a file system object is attempted, the list of
      rules is iterated until either a matching rule is located or
      the end is reached.  This behavior may be changed by the use
      of a &man.sysctl.8; parameter,
      <varname>security.mac.bsdextended.firstmatch_enabled</varname>.
      Similar to other firewall modules in &os;, a file containing
      the access control rules can be created and read by the system
      at boot time using an &man.rc.conf.5; variable.</para>

    <para>The rule list may be entered using &man.ugidfw.8; which has
      a syntax similar to &man.ipfw.8;.  More tools can be written by
      using the functions in the &man.libugidfw.3; library.</para>

    <para>Extreme caution should be taken when working with this
      module as incorrect use could block access to certain parts of
      the file system.</para>

    <sect2>
      <title>Examples</title>

      <para>After the &man.mac.bsdextended.4; module has been loaded,
	the following command may be used to list the current rule
	configuration:</para>

      <screen>&prompt.root; <userinput>ugidfw list</userinput>
0 slots, 0 rules</screen>

      <para>By default, no rules are defined and everything is
	completely accessible.  To create a rule which will block all
	access by users but leave <username>root</username>
	unaffected, run the following command:</para>

      <screen>&prompt.root; <userinput>ugidfw add subject not uid root new object not uid root mode n</userinput></screen>

      <para>This is a very bad idea as it will block all users from
	issuing even the most simple commands, such as
	<command>ls</command>.  The next example will block
	<username>user1</username> any and all access, including
	directory listings, to
	<username><replaceable>user2</replaceable></username>'s home
	directory:</para>

      <screen>&prompt.root; <userinput>ugidfw set 2 subject uid <replaceable>user1</replaceable> object uid <replaceable>user2</replaceable> mode n</userinput>
&prompt.root; <userinput>ugidfw set 3 subject uid <replaceable>user1</replaceable> object gid <replaceable>user2</replaceable> mode n</userinput></screen>

      <para>Instead of <username>user1</username>,
	<option>not uid <replaceable>user2</replaceable></option>
	could be used.  This enforces the same access restrictions for
	all users instead of just one user.</para>

      <note>
	<para>The <username>root</username> user is unaffected by
	  these changes.</para>
      </note>

      <para>For more information, refer to &man.mac.bsdextended.4; and
	&man.ugidfw.8;</para>
    </sect2>
  </sect1>

  <sect1 id="mac-ifoff">
    <title>The &man.mac.ifoff.4; Module</title>

    <indexterm>
      <primary>MAC Interface Silencing Policy</primary>
    </indexterm>
    <para>Module name: <filename>mac_ifoff.ko</filename></para>

    <para>Kernel configuration line:
      <literal>options MAC_IFOFF</literal></para>

    <para>Boot option:
      <literal>mac_ifoff_load="YES"</literal></para>

    <para>The &man.mac.ifoff.4; module exists solely to disable
      network interfaces on the fly and keep network interfaces from
      being brought up during system boot.  It does not require any
      labels to be set up on the system, nor does it depend on other
      <acronym>MAC</acronym> modules.</para>

    <para>Most of this module's control is performed through the
      <command>sysctl</command> tunables listed below.</para>

    <itemizedlist>
      <listitem>
	<para><varname>security.mac.ifoff.lo_enabled</varname>
	  enables or disables all traffic on the loopback (&man.lo.4;)
	  interface.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.ifoff.bpfrecv_enabled</varname>
	  enables or disables all traffic on the Berkeley Packet
	  Filter interface (&man.bpf.4;)</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.ifoff.other_enabled</varname>
	  enables or disables traffic on all other interfaces.</para>
      </listitem>
    </itemizedlist>

    <para>One of the most common uses of &man.mac.ifoff.4; is network
      monitoring in an environment where network traffic should not
      be permitted during the boot sequence.  Another suggested use
      would be to write a script which uses
      <filename role="package">security/aide</filename> to
      automatically block network traffic if it finds new or altered
      files in protected directories.</para>
  </sect1>

  <sect1 id="mac-portacl">
    <title>The &man.mac.portacl.4; Module</title>

    <indexterm>
      <primary>MAC Port Access Control List Policy</primary>
    </indexterm>
    <para>Module name: <filename>mac_portacl.ko</filename></para>

    <para>Kernel configuration line:
      <literal>MAC_PORTACL</literal></para>

    <para>Boot option:
      <literal>mac_portacl_load="YES"</literal></para>

    <para>The &man.mac.portacl.4; module is used to limit binding to
      local <acronym>TCP</acronym> and <acronym>UDP</acronym> ports
      using a variety of <command>sysctl</command> variables.
      &man.mac.portacl.4; makes it possible to allow
      non-<username>root</username> users to bind to specified
      privileged ports below 1024.</para>

    <para>Once loaded, this module enables the
      <acronym>MAC</acronym> policy on all sockets.  The following
      tunables are available:</para>

    <itemizedlist>
      <listitem>
	<para><varname>security.mac.portacl.enabled</varname>
	  enables or disables the policy completely.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.portacl.port_high</varname>
	  sets the highest port number that &man.mac.portacl.4;
	  protects.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.portacl.suser_exempt</varname>,
	  when set to a non-zero value, exempts the
	  <username>root</username> user from this policy.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.portacl.rules</varname>
	  specifies the mac_portacl policy, which is a text string of
	  the form: <literal>rule[,rule,...]</literal> with as many
	  rules as needed.  Each rule is of the form:
	  <literal>idtype:id:protocol:port</literal>.  The
	  <parameter>idtype</parameter> parameter can be
	  <literal>uid</literal> or <literal>gid</literal> and is used
	  to interpret the <parameter>id</parameter> parameter as
	  either a user id or group id, respectively.  The
	  <parameter>protocol</parameter> parameter is used to
	  determine if the rule should apply to <acronym>TCP</acronym>
	  or <acronym>UDP</acronym> by setting the parameter to
	  <literal>tcp</literal> or <literal>udp</literal>.  The final
	  <parameter>port</parameter> parameter is the port number to
	  allow the specified user or group to bind to.</para>
      </listitem>
    </itemizedlist>

    <note>
      <para>Since the ruleset is interpreted directly by the kernel,
	only numeric values can be used for the user ID, group ID,
	and port parameters.  Names cannot be used for users,
	groups, or services.</para>
    </note>

    <para>By default, ports below 1024 can only be used by or bound
      to privileged processes, which run as
      <username>root</username>.  For &man.mac.portacl.4; to allow
      non-privileged processes to bind to ports below 1024, this
      restriction has to be disabled by setting the &man.sysctl.8;
      variables
      <varname>net.inet.ip.portrange.reservedlow</varname> and
      <varname>net.inet.ip.portrange.reservedhigh</varname> to
      zero:</para>

    <screen>&prompt.root; <userinput>sysctl security.mac.portacl.port_high=1023</userinput>
&prompt.root; <userinput>sysctl net.inet.ip.portrange.reservedlow=0
net.inet.ip.portrange.reservedhigh=0</userinput></screen>

    <para>See the examples below or refer to &man.mac.portacl.4; for
      further information.</para>

    <sect2>
      <title>Examples</title>

      <para>Since the <username>root</username> user should not be
	crippled by this policy, this example starts by setting the
	<varname>security.mac.portacl.suser_exempt</varname> to a
	non-zero value.</para>

      <screen>&prompt.root; <userinput>sysctl security.mac.portacl.suser_exempt=1</userinput></screen>

      <para>Next, allow the user with <acronym>UID</acronym> 80
	to bind to port 80.  This allows the <username>www</username>
	user to run a web server without ever having
	<username>root</username> privilege.</para>

      <screen>&prompt.root; <userinput>sysctl security.mac.portacl.rules=uid:80:tcp:80</userinput></screen>

      <para>The next example permits the user with the
	<acronym>UID</acronym> of 1001 to bind to the
	<acronym>TCP</acronym> ports 110 (<quote>pop3</quote>) and 995
	(<quote>pop3s</quote>).  This permits this user to start a
	server that accepts connections on ports 110 and 995.</para>

      <screen>&prompt.root; <userinput>sysctl security.mac.portacl.rules=uid:1001:tcp:110,uid:1001:tcp:995</userinput></screen>

    </sect2>
  </sect1>

  <sect1 id="mac-partition">
    <title>The &man.mac.partition.4; Module</title>

    <indexterm>
      <primary>MAC Process Partition Policy</primary>
    </indexterm>
    <para>Module name: <filename>mac_partition.ko</filename></para>

    <para>Kernel configuration line:
      <literal>options MAC_PARTITION</literal></para>

    <para>Boot option:
      <literal>mac_partition_load="YES"</literal></para>

    <para>The &man.mac.partition.4; policy will drop processes into
      specific <quote>partitions</quote> based on their
      <acronym>MAC</acronym> label.  This module should be added to
      &man.loader.conf.5; so that it loads and enables the policy
      at system boot.</para>

    <para>Most configuration for this policy is done using
      &man.setpmac.8;.  One <command>sysctl</command> tunable is
      available for this policy:</para>

    <itemizedlist>
      <listitem>
	<para><varname>security.mac.partition.enabled</varname>
	  enables the enforcement of <acronym>MAC</acronym> process
	  partitions.</para>
      </listitem>
    </itemizedlist>

    <para>When this policy is enabled, users will only be permitted
      to see their processes, and any others within their partition,
      but will not be permitted to work with utilities outside the
      scope of this partition.  For instance, a user in the
      <literal>insecure</literal> class will not be permitted to
      access <command>top</command> as well as many other commands
      that must spawn a process.</para>

    <para>To set or drop utilities into a partition label, use the
      <command>setpmac</command> utility:</para>

    <screen>&prompt.root; <userinput>setpmac partition/13 top</userinput></screen>

    <para>This example adds <command>top</command> to the label set on
      users in the <literal>insecure</literal> class.  All processes
      spawned by users in the <literal>insecure</literal> class will
      stay in the <literal>partition/13</literal> label.</para>

    <sect2>
      <title>Examples</title>

      <para>The following command will display the partition label
	and the process list:</para>

      <screen>&prompt.root; <userinput>ps Zax</userinput></screen>

      <para>This command will display another user's process partition
	label and that user's currently running processes:</para>

      <screen>&prompt.root; <userinput>ps -ZU trhodes</userinput></screen>

      <note>
	<para>Users can see processes in <username>root</username>'s
	  label unless the &man.mac.seeotheruids.4; policy is
	  loaded.</para>
      </note>

      <para>A really crafty implementation could have all of the
	services disabled in <filename>/etc/rc.conf</filename> and
	started by a script that starts them with the proper
	labeling set.</para>

      <note>
	<para>The following policies support integer settings
	  in place of the three default labels offered.  These
	  options, including their limitations, are further explained
	  in the module manual pages.</para>
      </note>
    </sect2>
  </sect1>

  <sect1 id="mac-mls">
    <title>The MAC Multi-Level Security Module</title>

    <indexterm>
      <primary>MAC Multi-Level Security Policy</primary>
    </indexterm>
    <para>Module name: <filename>mac_mls.ko</filename></para>

    <para>Kernel configuration line:
      <literal>options MAC_MLS</literal></para>

    <para>Boot option: <literal>mac_mls_load="YES"</literal></para>

    <para>The &man.mac.mls.4; policy controls access between subjects
      and objects in the system by enforcing a strict information
      flow policy.</para>

    <para>In <acronym>MLS</acronym> environments, a
      <quote>clearance</quote> level is set in the label of each
      subject or object, along with compartments.  Since these
      clearance or sensibility levels can reach numbers greater than
      several thousand; it would be a daunting task for any system
      administrator to thoroughly configure each subject or object.
      Thankfully, three <quote>instant</quote> labels are included in
      this policy.</para>

    <para>These labels are <literal>mls/low</literal>,
      <literal>mls/equal</literal> and <literal>mls/high</literal>.
      Since these labels are described in depth in the manual page,
      they will only get a brief description here:</para>

    <itemizedlist>
      <listitem>
	<para>The <literal>mls/low</literal> label contains a low
	  configuration which permits it to be dominated by all other
	  objects.  Anything labeled with <literal>mls/low</literal>
	  will have a low clearance level and not be permitted to
	  access information of a higher level.  This label also
	  prevents objects of a higher clearance level from writing or
	  passing information on to them.</para>
      </listitem>

      <listitem>
	<para>The <literal>mls/equal</literal> label should be
	  placed on objects considered to be exempt from the
	  policy.</para>
      </listitem>

      <listitem>
	<para>The <literal>mls/high</literal> label is the highest
	  level of clearance possible.  Objects assigned this label
	  will hold dominance over all other objects in the system;
	  however, they will not permit the leaking of information
	  to objects of a lower class.</para>
      </listitem>
    </itemizedlist>

    <para><acronym>MLS</acronym> provides:</para>

    <itemizedlist>
      <listitem>
	<para>A hierarchical security level with a set of non
	  hierarchical categories.</para>
      </listitem>

      <listitem>
	<para>Fixed rules of <literal>no read up, no write
	    down</literal>.  This means that a subject can have read
	  access to objects on its own level or below, but not above.
	  Similarly, a subject can have write access to objects on its
	  own level or above but not beneath.</para>
      </listitem>

      <listitem>
	<para>Secrecy, or the prevention of inappropriate disclosure
	  of data.</para>
      </listitem>

      <listitem>
	<para>A basis for the design of systems that concurrently
	  handle data at multiple sensitivity levels without leaking
	  information between secret and confidential.</para>
      </listitem>
    </itemizedlist>

    <para>The following <command>sysctl</command> tunables are
      available for the configuration of special services and
      interfaces:</para>

    <itemizedlist>
      <listitem>
	<para><varname>security.mac.mls.enabled</varname> is used to
	  enable or disable the <acronym>MLS</acronym> policy.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.mls.ptys_equal</varname>
	  labels all &man.pty.4; devices as
	  <literal>mls/equal</literal> during creation.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.mls.revocation_enabled</varname>
	  revokes access to objects after their label changes to a
	  label of a lower grade.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.mls.max_compartments</varname>
	  sets the maximum number of compartment levels allowed on a
	  system.</para>
      </listitem>
    </itemizedlist>

    <para>To manipulate the <acronym>MLS</acronym> labels, use
      &man.setfmac.8;.  To assign a label to an object, issue the
      following command:</para>

    <screen>&prompt.root; <userinput>setfmac mls/5 test</userinput></screen>

    <para>To get the <acronym>MLS</acronym> label for the file
      <filename>test</filename>, issue the following command:</para>

    <screen>&prompt.root; <userinput>getfmac test</userinput></screen>

    <para>Another approach is to create a master policy file in
      <filename class="directory">/etc/</filename> which specifies the
      <acronym>MLS</acronym> policy information and to feed that file
      to <command>setfmac</command>.  This method will be explained
      after all policies are covered.</para>

    <sect2>
      <title>Planning Mandatory Sensitivity</title>

      <para>When using the MLS policy module, an administrator plans
	to control the flow of sensitive information.  The default
	<literal>block read up block write down</literal> sets
	everything to a low state.  Everything is accessible and an
	administrator slowly augments the confidentiality of the
	information during the configuration stage;.</para>

      <para>Beyond the three basic label options, an administrator may
	group users and groups as required to block the information
	flow between them.  It might be easier to look at the
	information in clearance levels using descriptive words, such
	as classifications of <literal>Confidential</literal>,
	<literal>Secret</literal>, and <literal>Top Secret</literal>.
	Some administrators instead create different groups based on
	project levels.  Regardless of the classification method, a
	well thought out plan must exist before implementing such a
	restrictive policy.</para>

      <para>Some example situations for the MLS policy module
	include an e-commerce web server, a file server holding
	critical company information, and financial institution
	environments.</para>
    </sect2>
  </sect1>

  <sect1 id="mac-biba">
    <title>The MAC Biba Module</title>

    <indexterm>
      <primary>MAC Biba Integrity Policy</primary>
    </indexterm>
    <para>Module name: <filename>mac_biba.ko</filename></para>

    <para>Kernel configuration line: <literal>options
	MAC_BIBA</literal></para>

    <para>Boot option: <literal>mac_biba_load="YES"</literal></para>

    <para>The &man.mac.biba.4; module loads the <acronym>MAC</acronym>
      Biba policy.  This policy is similar to the
      <acronym>MLS</acronym> policy with the exception that the rules
      for information flow are slightly reversed.  This is to prevent
      the downward flow of sensitive information whereas the
      <acronym>MLS</acronym> policy prevents the upward flow of
      sensitive information.  Much of this section can apply to both
      policies.</para>

    <para>In Biba environments, an <quote>integrity</quote> label is
      set on each subject or object.  These labels are made up of
      hierarchical grades and non-hierarchical components.  As an
      grade ascends, so does its integrity.</para>

    <para>Supported labels are <literal>biba/low</literal>,
      <literal>biba/equal</literal>, and <literal>biba/high</literal>;
      as explained below:</para>

    <itemizedlist>
      <listitem>
	<para>The <literal>biba/low</literal> label is considered the
	  lowest integrity an object or subject may have.  Setting
	  this on objects or subjects will block their write access
	  to objects or subjects marked high.  They still have read
	  access though.</para>
      </listitem>

      <listitem>
	<para>The <literal>biba/equal</literal> label should only be
	  placed on objects considered to be exempt from the
	  policy.</para>
      </listitem>

      <listitem>
	<para>The <literal>biba/high</literal> label will permit
	  writing to objects set at a lower label, but not
	  permit reading that object.  It is recommended that this
	  label be placed on objects that affect the integrity of
	  the entire system.</para>
      </listitem>
    </itemizedlist>

    <para>Biba provides:</para>

    <itemizedlist>
      <listitem>
	<para>Hierarchical integrity level with a set of non
	  hierarchical integrity categories.</para>
      </listitem>

      <listitem>
	<para>Fixed rules are <literal>no write up, no read
	    down</literal>, the opposite of
	  <acronym>MLS</acronym>.  A subject can have write access
	  to objects on its own level or below, but not above.
	  Similarly, a subject can have read access to objects on
	  its own level or above, but not below.</para>
      </listitem>

      <listitem>
	<para>Integrity by preventing inappropriate modification of
	  data.</para>
      </listitem>

      <listitem>
	<para>Integrity levels instead of MLS sensitivity
	  levels.</para>
      </listitem>
    </itemizedlist>

    <para>The following <command>sysctl</command> tunables can
      be used to manipulate the Biba policy:</para>

    <itemizedlist>
      <listitem>
	<para><varname>security.mac.biba.enabled</varname> is used
	  to enable or disable enforcement of the Biba policy on the
	  target machine.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.biba.ptys_equal</varname> is
	  used to disable the Biba policy on &man.pty.4;
	  devices.</para>
      </listitem>

      <listitem>
	<para><varname>security.mac.biba.revocation_enabled</varname>
	  forces the revocation of access to objects if the label
	  is changed to dominate the subject.</para>
      </listitem>
    </itemizedlist>

    <para>To access the Biba policy setting on system objects, use
      <command>setfmac</command> and
      <command>getfmac</command>:</para>

    <screen>&prompt.root; <userinput>setfmac biba/low test</userinput>
&prompt.root; <userinput>getfmac test</userinput>
test: biba/low</screen>

    <sect2>
      <title>Planning Mandatory Integrity</title>

      <para>Integrity, which is different from sensitivity, guarantees
	that the information will never be manipulated by untrusted
	parties.  This includes information passed between subjects,
	objects, and both.  It ensures that users will only be able to
	modify or access information they explicitly need to.</para>

      <para>The &man.mac.biba.4; security policy module permits an
	administrator to address which files and programs a user may
	see and invoke while assuring that the programs and files are
	free from threats and trusted by the system for that
	user.</para>

      <para>During the initial planning phase, an administrator must
	be prepared to partition users into grades, levels, and areas.
	Users will be blocked access not only to data but to programs
	and utilities both before and after they start.  The system
	will default to a high label once this policy module is
	enabled, and it is up to the administrator to configure the
	different grades and levels for users.  Instead of using
	clearance levels, a good planning method could include topics.
	For instance, only allow developers modification access to the
	source code repository, source code compiler, and other
	development utilities.  Other users would be grouped into
	other categories such as testers, designers, or end users and
	would only be permitted read access.</para>

      <para>A lower integrity subject is unable to write to a higher
	integrity subject and a higher integrity subject cannot
	observe or read a lower integrity object.  Setting a label at
	the lowest possible grade could make it inaccessible to
	subjects.  Some prospective environments for this security
	policy module would include a constrained web server, a
	development and test machine, and a source code repository.  A
	less useful implementation would be a personal workstation, a
	machine used as a router, or a network firewall.</para>
    </sect2>
  </sect1>

  <sect1 id="mac-lomac">
    <title>The MAC LOMAC Module</title>

    <indexterm>
      <primary>MAC LOMAC</primary>
    </indexterm>
    <para>Module name: <filename>mac_lomac.ko</filename></para>

    <para>Kernel configuration line: <literal>options
	MAC_LOMAC</literal></para>

    <para>Boot option: <literal>mac_lomac_load="YES"</literal></para>

    <para>Unlike the <acronym>MAC</acronym> Biba policy, the
      &man.mac.lomac.4; policy permits access to lower integrity
      objects only after decreasing the integrity level to not disrupt
      any integrity rules.</para>

    <para>The <acronym>MAC</acronym> version of the Low-watermark
      integrity policy works almost identically to Biba, but with the
      exception of using floating labels to support subject demotion
      via an auxiliary grade compartment.  This secondary compartment
      takes the form <literal>[auxgrade]</literal>.  When assigning
      a LOMAC policy with an auxiliary grade, use the syntax
      <literal>lomac/10[2]</literal> where the number two (2) is the
      auxiliary grade.</para>

    <para>The <acronym>MAC</acronym> LOMAC policy relies on the
      ubiquitous labeling of all system objects with integrity labels,
      permitting subjects to read from low integrity objects and then
      downgrading the label on the subject to prevent future writes to
      high integrity objects using <literal>[auxgrade]</literal>.  The
      policy may provide for greater compatibility and require less
      initial configuration than Biba.</para>

    <sect2>
      <title>Examples</title>

      <para>Like the Biba and <acronym>MLS</acronym> policies,
	<command>setfmac</command> and <command>setpmac</command>
	are used to place labels on system objects:</para>

      <screen>&prompt.root; <userinput>setfmac /usr/home/trhodes lomac/high[low]</userinput>
&prompt.root; <userinput>getfmac /usr/home/trhodes</userinput> lomac/high[low]</screen>

      <para>The auxiliary grade <literal>low</literal> is a feature
	provided only by the <acronym>MAC</acronym> LOMAC
	policy.</para>
    </sect2>
  </sect1>

  <sect1 id="mac-implementing">
    <title>Nagios in a MAC Jail</title>

    <indexterm>
      <primary>Nagios in a MAC Jail</primary>
    </indexterm>

    <para>The following demonstration implements a secure
      environment using various <acronym>MAC</acronym> modules
      with properly configured policies.  This is only a test as
      implementing a policy and ignoring it could be disastrous in a
      production environment.</para>

    <para>Before beginning this process, <option>multilabel</option>
      must be set on each file system as not doing so will result in
      errors.  This example assumes that <filename
	role="package">net-mngt/nagios-plugins</filename>,
      <filename role="package">net-mngt/nagios</filename>, and
      <filename role="package">www/apache22</filename> are all
      installed, configured, and working correctly.</para>

    <sect2>
      <title>Create an Insecure User Class</title>

      <para>Begin the procedure by adding the following user class
	to <filename>/etc/login.conf</filename>:</para>

      <programlisting>insecure:\
:copyright=/etc/COPYRIGHT:\
:welcome=/etc/motd:\
:setenv=MAIL=/var/mail/$,BLOCKSIZE=K:\
:path=~/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin
:manpath=/usr/share/man /usr/local/man:\
:nologin=/usr/sbin/nologin:\
:cputime=1h30m:\
:datasize=8M:\
:vmemoryuse=100M:\
:stacksize=2M:\
:memorylocked=4M:\
:memoryuse=8M:\
:filesize=8M:\
:coredumpsize=8M:\
:openfiles=24:\
:maxproc=32:\
:priority=0:\
:requirehome:\
:passwordtime=91d:\
:umask=022:\
:ignoretime@:\
:label=biba/10(10-10):</programlisting>

      <para>Add the following line to the default user class:</para>

      <programlisting>:label=biba/high:</programlisting>

      <para>Next, issue the following command to rebuild the
	database:</para>

      <screen>&prompt.root; <userinput>cap_mkdb /etc/login.conf</userinput></screen>
    </sect2>

    <sect2>
      <title>Boot Configuration</title>

      <para>Add the following lines to
	<filename>/boot/loader.conf</filename>:</para>

      <programlisting>mac_biba_load="YES"
mac_seeotheruids_load="YES"</programlisting>
    </sect2>

    <sect2>
      <title>Configure Users</title>

      <para>Set the <username>root</username> user to the default
	class using:</para>

      <screen>&prompt.root; <userinput>pw usermod root -L default</userinput></screen>

      <para>All user accounts that are not <username>root</username>
	or system users will now require a login class.  The login
	class is required otherwise users will be refused access
	to common commands such as &man.vi.1;.  The following
	<command>sh</command> script should do the trick:</para>

      <screen>&prompt.root; <userinput>for x in `awk -F: '($3 &gt;= 1001) &amp;&amp; ($3 != 65534) { print $1 }' \</userinput>
	<userinput>/etc/passwd`; do pw usermod $x -L default; done;</userinput></screen>

      <para>Drop the <username>nagios</username> and
	<username>www</username> users into the insecure class:</para>

      <screen>&prompt.root; <userinput>pw usermod nagios -L insecure</userinput></screen>

      <screen>&prompt.root; <userinput>pw usermod www -L insecure</userinput></screen>

      </sect2>

      <sect2>
	<title>Create the Contexts File</title>

	<para>A contexts file should now be created as
	  <filename>/etc/policy.contexts</filename>.</para>

	<programlisting># This is the default BIBA policy for this system.

# System:
/var/run                        biba/equal
/var/run/*                      biba/equal

/dev                            biba/equal
/dev/*                          biba/equal

/var				biba/equal
/var/spool                      biba/equal
/var/spool/*                    biba/equal

/var/log                        biba/equal
/var/log/*                      biba/equal

/tmp				biba/equal
/tmp/*				biba/equal
/var/tmp			biba/equal
/var/tmp/*			biba/equal

/var/spool/mqueue		biba/equal
/var/spool/clientmqueue		biba/equal

# For Nagios:
/usr/local/etc/nagios
/usr/local/etc/nagios/*         biba/10

/var/spool/nagios               biba/10
/var/spool/nagios/*             biba/10

# For apache
/usr/local/etc/apache           biba/10
/usr/local/etc/apache/*         biba/10</programlisting>

      <para>This policy enforces security by setting restrictions
	on the flow of information.  In this specific configuration,
	users, including <username>root</username>, should never be
	allowed to access <application>Nagios</application>.
	Configuration files and processes that are a part of
	<application>Nagios</application> will be completely self
	contained or jailed.</para>

      <para>This file will be read by the system by issuing the
	following command:</para>

      <screen>&prompt.root; <userinput>setfsmac -ef /etc/policy.contexts /</userinput>
&prompt.root; <userinput>setfsmac -ef /etc/policy.contexts /</userinput></screen>

      <note>
	<para>The above file system layout will differ depending
	  upon the environment and must be run on every file
	  system.</para>
      </note>

      <para><filename>/etc/mac.conf</filename> requires the following
	modifications in the main section:</para>

      <programlisting>default_labels file ?biba
default_labels ifnet ?biba
default_labels process ?biba
default_labels socket ?biba</programlisting>
    </sect2>

    <sect2>
      <title>Enable Networking</title>

      <para>Add the following line to
	<filename>/boot/loader.conf</filename>:</para>

      <programlisting>security.mac.biba.trust_all_interfaces=1</programlisting>

      <para>And the following to the network card configuration stored
	in <filename>rc.conf</filename>.  If the primary Internet
	configuration is done via <acronym>DHCP</acronym>, this may
	need to be configured manually after every system boot:</para>

      <programlisting>maclabel biba/equal</programlisting>
    </sect2>

    <sect2>
      <title>Testing the Configuration</title>

      <indexterm>
	<primary>MAC Configuration Testing</primary>
      </indexterm>

      <para>Ensure that the web server and
	<application>Nagios</application> will not be started on
	system initialization and reboot.  Ensure the
	<username>root</username> user cannot access any of the files
	in the <application>Nagios</application> configuration
	directory.  If <username>root</username> can issue an
	&man.ls.1; command on <filename>/var/spool/nagios</filename>,
	something is wrong.  Otherwise a <quote>permission
	  denied</quote> error should be returned.</para>

      <para>If all seems well, <application>Nagios</application>,
	<application>Apache</application>, and
	<application>Sendmail</application> can now be started:</para>

      <screen>&prompt.root; <userinput>cd /etc/mail &amp;&amp; make stop &amp;&amp; \
setpmac biba/equal make start &amp;&amp; setpmac biba/10\(10-10\) apachectl start &amp;&amp; \
setpmac biba/10\(10-10\) /usr/local/etc/rc.d/nagios.sh forcestart</userinput></screen>

      <para>Double check to ensure that everything is working
	properly.  If not, check the log files for error messages.
	Use &man.sysctl.8; to disable the &man.mac.biba.4; security
	policy module enforcement and try starting everything again as
	usual.</para>

      <note>
	<para>The <username>root</username> user can still change the
	  security enforcement and edit its configuration files.  The
	  following command will permit the degradation of the
	  security policy to a lower grade for a newly spawned
	  shell:</para>

	<screen>&prompt.root; <userinput>setpmac biba/10 csh</userinput></screen>

	<para>To block this from happening, force the user into a
	  range using &man.login.conf.5;.  If &man.setpmac.8; attempts
	  to run a command outside of the compartment's range, an
	  error will be returned and the command will not be executed.
	  In this case, set root to
	  <literal>biba/high(high-high)</literal>.</para>
      </note>
    </sect2>
  </sect1>

  <sect1 id="mac-userlocked">
    <title>User Lock Down</title>

    <para>This example considers a relatively small storage system
      with fewer than fifty users.  Users will have login
      capabilities, and be permitted to store data and access
      resources.</para>

    <para>For this scenario, the &man.mac.bsdextended.4; and
      &man.mac.seeotheruids.4; policy modules could co-exist and block
      access to system objects while hiding user processes.</para>

    <para>Begin by adding the following line to
      <filename>/boot/loader.conf</filename>:</para>

    <programlisting>mac_seeotheruids_load="YES"</programlisting>

    <para>The &man.mac.bsdextended.4; security policy module may be
      activated by adding this line to
      <filename>/etc/rc.conf</filename>:</para>

    <programlisting>ugidfw_enable="YES"</programlisting>

    <para>Default rules stored in
      <filename>/etc/rc.bsdextended</filename> will be loaded at
      system initialization.  However, the default entries may need
      modification.  Since this machine is expected only to service
      users, everything may be left commented out except the last
      two lines in order to force the loading of user owned system
      objects by default.</para>

    <para>Add the required users to this machine and reboot.  For
      testing purposes, try logging in as a different user across
      two consoles.  Run <command>ps aux</command> to see if processes
      of other users are visible.  Verify that running &man.ls.1; on
      another user's home directory fails.</para>

    <para>Do not try to test with the <username>root</username> user
      unless the specific <command>sysctl</command>s have been
      modified to block super user access.</para>

    <note>
      <para>When a new user is added, their &man.mac.bsdextended.4;
	rule will not be in the ruleset list.  To update the ruleset
	quickly, unload the security policy module and reload it again
	using &man.kldunload.8; and &man.kldload.8;.</para>
    </note>
  </sect1>

  <sect1 id="mac-troubleshoot">
    <title>Troubleshooting the MAC Framework</title>

    <indexterm>
      <primary>MAC Troubleshooting</primary>
    </indexterm>

    <para>This section discusses common configuration issues.</para>

    <sect2>
      <title><option>multilabel</option> cannot be enabled on
	<filename>/</filename></title>

      <para>The<option>multilabel</option> flag does not stay
	enabled on my root (<filename>/</filename>) partition!</para>


      <para>The following steps may resolve this transient
	error:</para>

      <procedure>
	<step>
	  <para>Edit <filename>/etc/fstab</filename> and set the root
	    partition to <option>ro</option> for read-only.</para>
	</step>

	<step>
	  <para>Reboot into single user mode.</para>
	</step>

	<step>
	  <para>Run <command>tunefs</command> <option>-l
	      enable</option>
	    on <filename>/</filename>.</para>
	</step>

	<step>
	  <para>Reboot the system.</para>
	</step>

	<step>
	  <para>Run <command>mount</command> <option>-urw</option>
	    <filename>/</filename> and change the <option>ro</option>
	    back to <option>rw</option> in
	    <filename>/etc/fstab</filename> and reboot the system
	    again.</para>
	</step>

	<step>
	  <para>Double-check the output from
	    <command>mount</command> to ensure that
	    <option>multilabel</option> has been properly set on the
	    root file system.</para>
	</step>
      </procedure>
    </sect2>

    <sect2>
      <title>Xorg Server Will Not Start After
	<acronym>MAC</acronym></title>

      <para>After establishing a secure environment with
	<acronym>MAC</acronym>, I am no longer able to start
	Xorg!</para>

      <para>This could be caused by the <acronym>MAC</acronym>
	<literal>partition</literal> policy or by a mislabeling in
	one of the <acronym>MAC</acronym> labeling policies.  To
	debug, try the following:</para>

      <procedure>
	<step>
	  <para>Check the error message; if the user is in the
	    <literal>insecure</literal> class, the
	    <literal>partition</literal> policy may be the culprit.
	    Try setting the user's class back to the
	    <literal>default</literal> class and rebuild the database
	    with <command>cap_mkdb</command>.  If this does not
	    alleviate the problem, go to step two.</para>
	</step>

	<step>
	  <para>Double-check the label policies.  Ensure that the
	    policies are set correctly for the user, the Xorg
	    application, and the <filename
	      class="directory">/dev</filename> entries.</para>
	</step>

	<step>
	  <para>If neither of these resolve the problem, send the
	    error message and a description of the environment to
	    the &a.questions; mailing list.</para>
	</step>
      </procedure>
    </sect2>

    <sect2>
      <title>Error: &man..secure.path.3; cannot stat
	<filename>.login_conf</filename></title>

      <para>When a user attempts to switch from the
	<username>root</username> user to another user in the system,
	the error message <errorname>_secure_path: unable to state
	  .login_conf</errorname> appears.</para>

      <para>This message is usually shown when the user has a higher
	label setting than that of the user they are attempting to
	become.  For instance, <username>joe</username> has a default
	label of <option>biba/low</option>.  The
	<username>root</username> user, who has a label of
	<option>biba/high</option>, cannot view
	<username>joe</username>'s home directory.  This will happen
	whether or not <username>root</username> has used
	<command>su</command> to become <username>joe</username> as
	the Biba integrity model will not permit
	<username>root</username> to view objects set at a lower
	integrity level.</para>
    </sect2>

    <sect2>
      <title>The <username>root</username> username is broken!</title>

      <para>In normal or even single user mode, the
	<username>root</username> is not recognized,
	<command>whoami</command> returns 0 (zero), and
	<command>su</command> returns <errorname>who are
	  you?</errorname>.</para>

      <para>This can happen if a labeling policy has been disabled,
	either by a &man.sysctl.8; or the policy module was unloaded.
	If the policy is disabled, the login capabilities database
	needs to be reconfigured with <option>label</option> removed.
	Double check <filename>login.conf</filename> to ensure that
	all <option>label</option> options have been removed and
	rebuild the database with <command>cap_mkdb</command>.</para>

      <para>This may also happen if a policy restricts access to
	<filename>master.passwd</filename>.  This is usually caused by
	an administrator altering the file under a label which
	conflicts with the general policy being used by the system.
	In these cases, the user information would be read by the
	system and access would be blocked as the file has inherited
	the new label.  Disable the policy using &man.sysctl.8; and
	everything should return to normal.</para>
    </sect2>
  </sect1>
</chapter>