apparmor/parser/apparmor_parser.pod

# ----------------------------------------------------------------------
#    Copyright (c) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007,
#                  2008, 2009
#    NOVELL (All rights reserved)
#
#    Copyright (c) 2010 - 2012
#    Canonical Ltd. (All rights reserved)
#
#    This program is free software; you can redistribute it and/or
#    modify it under the terms of version 2 of the GNU General Public
#    License published by the Free Software Foundation.
#
#    This program is distributed in the hope that it will be useful,
#    but WITHOUT ANY WARRANTY; without even the implied warranty of
#    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#    GNU General Public License for more details.
#
#    You should have received a copy of the GNU General Public License
#    along with this program; if not, contact Novell, Inc.
# ----------------------------------------------------------------------


=pod

=head1 NAME

apparmor_parser - loads AppArmor profiles into the kernel

=head1 SYNOPSIS

B<apparmor_parser [options] E<lt>commandE<gt> [profiles]...>

B<apparmor_parser [options] E<lt>commandE<gt>>

B<apparmor_parser [-hv] [--help] [--version]>

=head1 DESCRIPTION

B<apparmor_parser> is used as a general tool to compile, and manage AppArmor
policy, including loading new apparmor.d(5) profiles into the Linux kernel.

AppArmor profiles restrict the operations available to processes.

The B<profiles> are loaded into the Linux kernel by the B<apparmor_parser>
program. The B<profiles> may be specified by file name or a directory
name containing a set of profiles. If a directory is specified then the
B<apparmor_parser> will try to do a profile load for each file in the
directory that is not a dot file, or explicitly black listed (*.dpkg-new,
*.dpkg-old, *.dpkg-dist, *.dpkg-bak, *.dpkg-remove, *.pacsave, *.pacnew,
*.rpmnew, *.rpmsave, *.orig, *.rej, *~).
The B<apparmor_parser> will fall back to taking input from standard input if
a profile or directory is not supplied.

The input supplied to B<apparmor_parser> should be in the format described in
apparmor.d(5).

=head1 COMMANDS

The command set is broken into four subcategories.

=over 4

=item unprivileged commands

Commands that don't require any privilege and don't operate on profiles.

=item unprivileged profile commands

Commands that operate on a profile either specified on the command line or
read from stdin if no profile was specified.

=item privileged commands

Commands that require the MAC_ADMIN capability within the affected AppArmor
namespace to load policy into the kernel or filesystem write permissions to
update the affected privileged files (cache etc).

=item privileged profile commands

Commands that require privilege and operate on profiles.

=back

=head1 Unprivileged commands

=over 4

=item -V, --version

Print the version number and exit.

=item -h, --help

Give a quick reference guide.

=back

=head1 Unprivileged profile commands

=over 4

=item -N, --names

Produce a list of policies from a given set of profiles (implies -K).

=item -p, --preprocess

Apply preprocessing to the input profile(s) by flattening includes into
the output profile and dump to stdout.

=item -S, --stdout

Writes a binary (cached) profile to stdout (implies -K and -T).

=item -o file, --ofile file

Writes a binary (cached) profile to the specified file (implies -K and -T)

=back

=head1 Privileged commands

=over 4

=item --purge-cache

Unconditionally clear out cached profiles.

=back

=head1 Privileged profile commands

=over 4

=item -a, --add

Insert the AppArmor definitions given into the kernel. This is the default
action. This gives an error message if a AppArmor definition by the same
name already exists in the kernel, or if the parser doesn't understand
its input. It reports when an addition succeeded.

=item -r, --replace

This flag is required if an AppArmor definition by the same name already
exists in the kernel; used to replace the definition already
in the kernel with the definition given on standard input.

=item -R, --remove

This flag is used to remove an AppArmor definition already in the kernel.
Note that it still requires a complete AppArmor definition as described
in apparmor.d(5) even though the contents of the definition aren't
used.

=back

=head1 OPTIONS

=over 4

=item -B, --binary

Treat the profile files specified on the command line (or stdin if none
specified) as binary cache files, produced with the -S or -o options,
and load to the kernel as specified by -a, -r, and -R (implies -K
and -T).

=item -C, --Complain

Force the profile to load in complain mode.

=item -b n, --base n

Set the base directory for resolving #include directives
defined as relative paths.

=item -I n, --Include n

Add element n to the search path when resolving #include directives
defined as an absolute paths.

=item -f n, --subdomainfs n

Set the location of the apparmor security filesystem (default is
"/sys/kernel/security/apparmor").

=item -M n, --features-file n

Use the features file located at path "n" (default is
/etc/apparmor.d/cache/.features). If the --cache-loc option is present, the
".features" file in the specified cache directory is used.

=item -m n, --match-string n

Only use match features "n".

=item -n n, --namespace-string n

Force a profile to load in the namespace "n".

=item -X, --readimpliesX

In the case of profiles that are loading on systems were READ_IMPLIES_EXEC
is set in the kernel for a given process, load the profile so that any "r"
flags are processed as "mr".

=item -k, --show-cache

Report the cache processing (hit/miss details) when loading or saving
cached profiles.

=item -K, --skip-cache

Perform no caching at all: disables -W, implies -T.

=item -T, --skip-read-cache

By default, if a profile's cache is found in the location specified by
--cache-loc and the timestamp is newer than the profile, it will be loaded
from the cache. This option disables this cache loading behavior.

=item -W, --write-cache

Write out cached profiles to the location specified in --cache-loc.  Off
by default. In cases where abstractions have been changed, and the parser
is running with "--replace", it may make sense to also use
"--skip-read-cache" with the "--write-cache" option.

=item --skip-bad-cache

Skip updating the cache if it contains cached profiles in a bad or
inconsistent state

=item -L, --cache-loc

Set the location(s) of the cache directory. This option can accept a
comma separated list of directories, which will be searched in order
to find a matching cache. The first matching cache file found is used
even if a directory later in the search order may contain a newer cache
file.

If multiple directories are specified and --write-cache has been specified
then cache writes will be made to the first directory in the list, all
other directories will be treated as read only.

If a cache directory name needs to have a comma as part of the name, it
can be specified by using a backslash to escape the comma character in
the directory name.

If not specified the cache location defaults to /etc/apparmor.d/cache.d

=item --print-cache-dir

Print the cache directory location. This path will be a subdirectory of the
directory specified by --cache-loc. The subdirectory used will be influenced by
the features available in the currently running kernel or by the features
specified with the --match-string or --features-file options.

=item -Q, --skip-kernel-load

Perform all actions except the actual loading of a profile into the kernel.
This is useful for testing profile generation, caching, etc, without making
changes to the running kernel profiles.

This also removes the need for privilege to execute the commands that
manage policy in the kernel

=item -q, --quiet

Do not report on the profiles as they are loaded, and not show warnings.

=item -v, --verbose

Report on the profiles as they are loaded, and show warnings.

=item --warn=n

Enable various warnings during policy compilation. A single dump flag
can be specified per --warn option, but the --warn flag can be passed
multiple times.

  apparmor_parser --warn=rules-not-enforced ...

Use --help=warn to see a full list of which warn flags are supported.

=item -d, --debug

Given once, only checks the profiles to ensure syntactic correctness.
Given twice, dumps its interpretation of the profile for checking.

=item -D n, --dump=n

Debug flag for dumping various structures and passes of policy compilation.
A single dump flag can be specified per --dump option, but the dump flag
can be passed multiple times.  Note progress flags tend to also imply
the matching stats flag.

  apparmor_parser --dump=dfa-stats --dump=trans-stats <file>

Use --help=dump to see a full list of which dump flags are supported

=item -j n, --jobs=n

Set the number of jobs used to compile the specified policy. Where n can
be

  0    - disable jobs and use the main process for all compilation
  #    - a specific number of jobs
  auto - the # of cpus in the in the system
  x#   - # * number of cpus

Eg.
  -j8     OR --jobs=8                   allows for 8 parallel jobs
  -jauto  OR --jobs=auto                sets the jobs to the # of cpus
  -jx4    OR --jobs=x4                  sets the jobs to # of cpus * 4
  -jx1   is equivalent to   -jauto

The default value is the number of cpus in the system.

=item --max-jobs n

Set a hard cap on the value that can be specified by the --jobs flag.
It takes the same set of options available to the --jobs option, and
defaults to 8*cpus

=item -O n, --optimize=n

Set the optimization flags used by policy compilation.  A single optimization
flag can be toggled per -O option, but the optimize flag can be passed
multiple times.  Turning off some phases of the optimization can make
it so that policy can't complete compilation due to size constraints
(it is entirely possible to create a dfa with millions of states that will
take days or longer to compile).

Note: The parser is set to use a balanced default set of flags, that
will result in reasonable compression but not take excessive amounts
of time to complete.

Use --help=optimize to see a full list of which optimization flags are
supported.

=item --abort-on-error
Abort processing of profiles on the first error encountered, otherwise
the parser will continue to try to compile other profiles if specified.

Note: If an error is encountered while processing profiles the last error
encountered will be used to set the exit code. 

=item --skip-bad-cache-rebuild
The default behavior of the parser is to check if a cached version of
a profile exists and if it does it attempt to load it into the kernel.
If that load is rejected, then the parser will attempt to rebuild the
cache file, and load again.

This option tells the parser to not attempt to rebuild the cache on
failure, instead the parser continues on with processing the remaining
profiles.

=back

=head1 CONFIG FILE

An optional config file /etc/apparmor/parser.conf can be used to specify the
default options for the parser, which then can be overridden using the command
line options.

The config file ignores leading whitespace and treats lines that begin with #
as comments.  Config options are specified one per line using the same format
as the longform command line options (without the preceding --).

Eg.
    #comment

    optimize=no-expr-tree
    optimize=compress-fast

As with the command line some options accumulate and others override, ie. when
there are conflicting versions of switch the last option is the one chosen.

Eg.
    Optimize=no-minimize
    Optimize=minimize

would result in Optimize=minimize being set.

The Include, Dump, and Optimize options accululate except for the inversion
option (no-X vs. X), and a couple options that work by setting/clearing
multiple options (compress-small).  In that case the option will override
the flags it sets but will may accumulate with others.

All other options override previously set values.

=head1 BUGS

If you find any bugs, please report them at
L<https://bugs.launchpad.net/apparmor/+filebug>.

=head1 SEE ALSO

apparmor(7), apparmor.d(5), subdomain.conf(5), aa_change_hat(2), and
L<http://wiki.apparmor.net>.

=cut