ansible/lib/ansible/module_utils/basic.py

# This code is part of Ansible, but is an independent component.
# This particular file snippet, and this file snippet only, is BSD licensed.
# Modules you write using this snippet, which is embedded dynamically by Ansible
# still belong to the author of the module, and may assign their own license
# to the complete work.
#
# Copyright (c), Michael DeHaan <michael.dehaan@gmail.com>, 2012-2013
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without modification,
# are permitted provided that the following conditions are met:
#
#    * Redistributions of source code must retain the above copyright
#      notice, this list of conditions and the following disclaimer.
#    * Redistributions in binary form must reproduce the above copyright notice,
#      this list of conditions and the following disclaimer in the documentation
#      and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
# IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
# USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#

BOOLEANS_TRUE = ['yes', 'on', '1', 'true', 1, True]
BOOLEANS_FALSE = ['no', 'off', '0', 'false', 0, False]
BOOLEANS = BOOLEANS_TRUE + BOOLEANS_FALSE

# ansible modules can be written in any language.  To simplify
# development of Python modules, the functions available here can
# be used to do many common tasks

import locale
import os
import re
import pipes
import shlex
import subprocess
import sys
import types
import time
import select
import shutil
import stat
import tempfile
import traceback
import grp
import pwd
import platform
import errno
import datetime
from itertools import repeat, chain

try:
    import syslog
    HAS_SYSLOG=True
except ImportError:
    HAS_SYSLOG=False

try:
    # Python 2
    from itertools import imap
except ImportError:
    # Python 3
    imap = map

try:
    # Python 2
    basestring
except NameError:
    # Python 3
    basestring = str

try:
    # Python 2
    unicode
except NameError:
    # Python 3
    unicode = str

try:
    # Python 2.6+
    bytes
except NameError:
    # Python 2.4
    bytes = str

try:
    dict.iteritems
except AttributeError:
    # Python 3
    def iteritems(d):
        return d.items()
else:
    # Python 2
    def iteritems(d):
        return d.iteritems()

try:
    reduce
except NameError:
    # Python 3
    from functools import reduce

try:
    NUMBERTYPES = (int, long, float)
except NameError:
    # Python 3
    NUMBERTYPES = (int, float)

# Python2 & 3 way to get NoneType
NoneType = type(None)

try:
    from collections import Sequence, Mapping
except ImportError:
    # python2.5
    Sequence = (list, tuple)
    Mapping = (dict,)

try:
    from collections.abc import KeysView
    SEQUENCETYPE = (Sequence, KeysView)
except:
    SEQUENCETYPE = Sequence

try:
    import json
    # Detect the python-json library which is incompatible
    # Look for simplejson if that's the case
    try:
        if not isinstance(json.loads, types.FunctionType) or not isinstance(json.dumps, types.FunctionType):
            raise ImportError
    except AttributeError:
        raise ImportError
except ImportError:
    try:
        import simplejson as json
    except ImportError:
        print('\n{"msg": "Error: ansible requires the stdlib json or simplejson module, neither was found!", "failed": true}')
        sys.exit(1)
    except SyntaxError:
        print('\n{"msg": "SyntaxError: probably due to installed simplejson being for a different python version", "failed": true}')
        sys.exit(1)

from ansible.module_utils.six import PY2, PY3, b, binary_type, text_type, string_types

HAVE_SELINUX=False
try:
    import selinux
    HAVE_SELINUX=True
except ImportError:
    pass

try:
    from systemd import journal
    has_journal = True
except ImportError:
    has_journal = False

AVAILABLE_HASH_ALGORITHMS = dict()
try:
    import hashlib

    # python 2.7.9+ and 2.7.0+
    for attribute in ('available_algorithms', 'algorithms'):
        algorithms = getattr(hashlib, attribute, None)
        if algorithms:
            break
    if algorithms is None:
        # python 2.5+
        algorithms = ('md5', 'sha1', 'sha224', 'sha256', 'sha384', 'sha512')
    for algorithm in algorithms:
        AVAILABLE_HASH_ALGORITHMS[algorithm] = getattr(hashlib, algorithm)
except ImportError:
    import sha
    AVAILABLE_HASH_ALGORITHMS = {'sha1': sha.sha}
    try:
        import md5
        AVAILABLE_HASH_ALGORITHMS['md5'] = md5.md5
    except ImportError:
        pass

try:
    from ast import literal_eval
except ImportError:
    # a replacement for literal_eval that works with python 2.4. from:
    # https://mail.python.org/pipermail/python-list/2009-September/551880.html
    # which is essentially a cut/paste from an earlier (2.6) version of python's
    # ast.py
    from compiler import ast, parse

    def literal_eval(node_or_string):
        """
        Safely evaluate an expression node or a string containing a Python
        expression.  The string or node provided may only consist of the  following
        Python literal structures: strings, numbers, tuples, lists, dicts,  booleans,
        and None.
        """
        _safe_names = {'None': None, 'True': True, 'False': False}
        if isinstance(node_or_string, basestring):
            node_or_string = parse(node_or_string, mode='eval')
        if isinstance(node_or_string, ast.Expression):
            node_or_string = node_or_string.node

        def _convert(node):
            if isinstance(node, ast.Const) and isinstance(node.value, (basestring, int, float, long, complex)):
                return node.value
            elif isinstance(node, ast.Tuple):
                return tuple(map(_convert, node.nodes))
            elif isinstance(node, ast.List):
                return list(map(_convert, node.nodes))
            elif isinstance(node, ast.Dict):
                return dict((_convert(k), _convert(v)) for k, v in node.items())
            elif isinstance(node, ast.Name):
                if node.name in _safe_names:
                    return _safe_names[node.name]
            elif isinstance(node, ast.UnarySub):
                return -_convert(node.expr)
            raise ValueError('malformed string')
        return _convert(node_or_string)

_literal_eval = literal_eval

# Backwards compat.  There were present in basic.py before
from ansible.module_utils.pycompat24 import get_exception

# Internal global holding passed in params.  This is consulted in case
# multiple AnsibleModules are created.  Otherwise each AnsibleModule would
# attempt to read from stdin.  Other code should not use this directly as it
# is an internal implementation detail
_ANSIBLE_ARGS = None

FILE_COMMON_ARGUMENTS=dict(
    src = dict(),
    mode = dict(type='raw'),
    owner = dict(),
    group = dict(),
    seuser = dict(),
    serole = dict(),
    selevel = dict(),
    setype = dict(),
    follow = dict(type='bool', default=False),
    # not taken by the file module, but other modules call file so it must ignore them.
    content = dict(no_log=True),
    backup = dict(),
    force = dict(),
    remote_src = dict(), # used by assemble
    regexp = dict(), # used by assemble
    delimiter = dict(), # used by assemble
    directory_mode = dict(), # used by copy
)

PASSWD_ARG_RE = re.compile(r'^[-]{0,2}pass[-]?(word|wd)?')

# Can't use 07777 on Python 3, can't use 0o7777 on Python 2.4
PERM_BITS = int('07777', 8)      # file mode permission bits
EXEC_PERM_BITS = int('00111', 8) # execute permission bits
DEFAULT_PERM = int('0666', 8)    # default file permission bits


def get_platform():
    ''' what's the platform?  example: Linux is a platform. '''
    return platform.system()

def get_distribution():
    ''' return the distribution name '''
    if platform.system() == 'Linux':
        try:
            supported_dists = platform._supported_dists + ('arch',)
            distribution = platform.linux_distribution(supported_dists=supported_dists)[0].capitalize()
            if not distribution and os.path.isfile('/etc/system-release'):
                distribution = platform.linux_distribution(supported_dists=['system'])[0].capitalize()
                if 'Amazon' in distribution:
                    distribution = 'Amazon'
                else:
                    distribution = 'OtherLinux'
        except:
            # FIXME: MethodMissing, I assume?
            distribution = platform.dist()[0].capitalize()
    else:
        distribution = None
    return distribution

def get_distribution_version():
    ''' return the distribution version '''
    if platform.system() == 'Linux':
        try:
            distribution_version = platform.linux_distribution()[1]
            if not distribution_version and os.path.isfile('/etc/system-release'):
                distribution_version = platform.linux_distribution(supported_dists=['system'])[1]
        except:
            # FIXME: MethodMissing, I assume?
            distribution_version = platform.dist()[1]
    else:
        distribution_version = None
    return distribution_version

def get_all_subclasses(cls):
    '''
    used by modules like Hardware or Network fact classes to retrieve all subclasses of a given class.
    __subclasses__ return only direct sub classes. This one go down into the class tree.
    '''
    # Retrieve direct subclasses
    subclasses = cls.__subclasses__()
    to_visit = list(subclasses)
    # Then visit all subclasses
    while to_visit:
        for sc in to_visit:
            # The current class is now visited, so remove it from list
            to_visit.remove(sc)
            # Appending all subclasses to visit and keep a reference of available class
            for ssc in sc.__subclasses__():
                subclasses.append(ssc)
                to_visit.append(ssc)
    return subclasses


def load_platform_subclass(cls, *args, **kwargs):
    '''
    used by modules like User to have different implementations based on detected platform.  See User
    module for an example.
    '''

    this_platform = get_platform()
    distribution = get_distribution()
    subclass = None

    # get the most specific superclass for this platform
    if distribution is not None:
        for sc in get_all_subclasses(cls):
            if sc.distribution is not None and sc.distribution == distribution and sc.platform == this_platform:
                subclass = sc
    if subclass is None:
        for sc in get_all_subclasses(cls):
            if sc.platform == this_platform and sc.distribution is None:
                subclass = sc
    if subclass is None:
        subclass = cls

    return super(cls, subclass).__new__(subclass)


def json_dict_unicode_to_bytes(d, encoding='utf-8'):
    ''' Recursively convert dict keys and values to byte str

        Specialized for json return because this only handles, lists, tuples,
        and dict container types (the containers that the json module returns)
    '''

    if isinstance(d, unicode):
        return d.encode(encoding)
    elif isinstance(d, dict):
        return dict(imap(json_dict_unicode_to_bytes, iteritems(d), repeat(encoding)))
    elif isinstance(d, list):
        return list(imap(json_dict_unicode_to_bytes, d, repeat(encoding)))
    elif isinstance(d, tuple):
        return tuple(imap(json_dict_unicode_to_bytes, d, repeat(encoding)))
    else:
        return d

def json_dict_bytes_to_unicode(d, encoding='utf-8'):
    ''' Recursively convert dict keys and values to byte str

        Specialized for json return because this only handles, lists, tuples,
        and dict container types (the containers that the json module returns)
    '''

    if isinstance(d, bytes):
        return unicode(d, encoding)
    elif isinstance(d, dict):
        return dict(imap(json_dict_bytes_to_unicode, iteritems(d), repeat(encoding)))
    elif isinstance(d, list):
        return list(imap(json_dict_bytes_to_unicode, d, repeat(encoding)))
    elif isinstance(d, tuple):
        return tuple(imap(json_dict_bytes_to_unicode, d, repeat(encoding)))
    else:
        return d

def return_values(obj):
    """ Return stringified values from datastructures. For use with removing
    sensitive values pre-jsonification."""
    if isinstance(obj, basestring):
        if obj:
            if isinstance(obj, bytes):
                yield obj
            else:
                # Unicode objects should all convert to utf-8
                # (still must deal with surrogateescape on python3)
                yield obj.encode('utf-8')
        return
    elif isinstance(obj, SEQUENCETYPE):
        for element in obj:
            for subelement in return_values(element):
                yield subelement
    elif isinstance(obj, Mapping):
        for element in obj.items():
            for subelement in return_values(element[1]):
                yield subelement
    elif isinstance(obj, (bool, NoneType)):
        # This must come before int because bools are also ints
        return
    elif isinstance(obj, NUMBERTYPES):
        yield str(obj)
    else:
        raise TypeError('Unknown parameter type: %s, %s' % (type(obj), obj))

def remove_values(value, no_log_strings):
    """ Remove strings in no_log_strings from value.  If value is a container
    type, then remove a lot more"""
    if isinstance(value, basestring):
        if isinstance(value, unicode):
            # This should work everywhere on python2. Need to check
            # surrogateescape on python3
            bytes_value = value.encode('utf-8')
            value_is_unicode = True
        else:
            bytes_value = value
            value_is_unicode = False
        if bytes_value in no_log_strings:
            return 'VALUE_SPECIFIED_IN_NO_LOG_PARAMETER'
        for omit_me in no_log_strings:
            bytes_value = bytes_value.replace(omit_me, '*' * 8)
        if value_is_unicode:
            value = unicode(bytes_value, 'utf-8', errors='replace')
        else:
            value = bytes_value
    elif isinstance(value, SEQUENCETYPE):
        return [remove_values(elem, no_log_strings) for elem in value]
    elif isinstance(value, Mapping):
        return dict((k, remove_values(v, no_log_strings)) for k, v in value.items())
    elif isinstance(value, tuple(chain(NUMBERTYPES, (bool, NoneType)))):
        stringy_value = str(value)
        if stringy_value in no_log_strings:
            return 'VALUE_SPECIFIED_IN_NO_LOG_PARAMETER'
        for omit_me in no_log_strings:
            if omit_me in stringy_value:
                return 'VALUE_SPECIFIED_IN_NO_LOG_PARAMETER'
    elif isinstance(value, datetime.datetime):
        value = value.isoformat()
    else:
        raise TypeError('Value of unknown type: %s, %s' % (type(value), value))
    return value


def heuristic_log_sanitize(data, no_log_values=None):
    ''' Remove strings that look like passwords from log messages '''
    # Currently filters:
    # user:pass@foo/whatever and http://username:pass@wherever/foo
    # This code has false positives and consumes parts of logs that are
    # not passwds

    # begin: start of a passwd containing string
    # end: end of a passwd containing string
    # sep: char between user and passwd
    # prev_begin: where in the overall string to start a search for
    #   a passwd
    # sep_search_end: where in the string to end a search for the sep
    output = []
    begin = len(data)
    prev_begin = begin
    sep = 1
    while sep:
        # Find the potential end of a passwd
        try:
            end = data.rindex('@', 0, begin)
        except ValueError:
            # No passwd in the rest of the data
            output.insert(0, data[0:begin])
            break

        # Search for the beginning of a passwd
        sep = None
        sep_search_end = end
        while not sep:
            # URL-style username+password
            try:
                begin = data.rindex('://', 0, sep_search_end)
            except ValueError:
                # No url style in the data, check for ssh style in the
                # rest of the string
                begin = 0
            # Search for separator
            try:
                sep = data.index(':', begin + 3, end)
            except ValueError:
                # No separator; choices:
                if begin == 0:
                    # Searched the whole string so there's no password
                    # here.  Return the remaining data
                    output.insert(0, data[0:begin])
                    break
                # Search for a different beginning of the password field.
                sep_search_end = begin
                continue
        if sep:
            # Password was found; remove it.
            output.insert(0, data[end:prev_begin])
            output.insert(0, '********')
            output.insert(0, data[begin:sep + 1])
            prev_begin = begin

    output = ''.join(output)
    if no_log_values:
        output = remove_values(output, no_log_values)
    return output

def is_executable(path):
    '''is the given path executable?

    Limitations:
    * Does not account for FSACLs.
    * Most times we really want to know "Can the current user execute this
      file"  This function does not tell us that, only if an execute bit is set.
    '''
    # These are all bitfields so first bitwise-or all the permissions we're
    # looking for, then bitwise-and with the file's mode to determine if any
    # execute bits are set.
    return ((stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH) & os.stat(path)[stat.ST_MODE])

def _load_params():
    ''' read the modules parameters and store them globally.

    This function may be needed for certain very dynamic custom modules which
    want to process the parameters that are being handed the module.  Since
    this is so closely tied to the implementation of modules we cannot
    guarantee API stability for it (it may change between versions) however we
    will try not to break it gratuitously.  It is certainly more future-proof
    to call this function and consume its outputs than to implement the logic
    inside it as a copy in your own code.
    '''
    global _ANSIBLE_ARGS
    if _ANSIBLE_ARGS is not None:
        buffer = _ANSIBLE_ARGS
    else:
        # debug overrides to read args from file or cmdline

        # Avoid tracebacks when locale is non-utf8
        # We control the args and we pass them as utf8
        if len(sys.argv) > 1:
            if os.path.isfile(sys.argv[1]):
                fd = open(sys.argv[1], 'rb')
                buffer = fd.read()
                fd.close()
            else:
                buffer = sys.argv[1]
                if PY3:
                    buffer = buffer.encode('utf-8', errors='surrogateescape')
        # default case, read from stdin
        else:
            if PY2:
                buffer = sys.stdin.read()
            else:
                buffer = sys.stdin.buffer.read()
        _ANSIBLE_ARGS = buffer

    try:
        params = json.loads(buffer.decode('utf-8'))
    except ValueError:
        # This helper used too early for fail_json to work.
        print('\n{"msg": "Error: Module unable to decode valid JSON on stdin.  Unable to figure out what parameters were passed", "failed": true}')
        sys.exit(1)

    if PY2:
        params = json_dict_unicode_to_bytes(params)

    try:
        return params['ANSIBLE_MODULE_ARGS']
    except KeyError:
        # This helper does not have access to fail_json so we have to print
        # json output on our own.
        print('\n{"msg": "Error: Module unable to locate ANSIBLE_MODULE_ARGS in json data from stdin.  Unable to figure out what parameters were passed", "failed": true}')
        sys.exit(1)

def env_fallback(*args, **kwargs):
    ''' Load value from environment '''
    for arg in args:
        if arg in os.environ:
            return os.environ[arg]
    else:
        raise AnsibleFallbackNotFound


class AnsibleFallbackNotFound(Exception):
    pass


class AnsibleModule(object):
    def __init__(self, argument_spec, bypass_checks=False, no_log=False,
        check_invalid_arguments=True, mutually_exclusive=None, required_together=None,
        required_one_of=None, add_file_common_args=False, supports_check_mode=False,
        required_if=None):

        '''
        common code for quickly building an ansible module in Python
        (although you can write modules in anything that can return JSON)
        see library/* for examples
        '''

        self.argument_spec = argument_spec
        self.supports_check_mode = supports_check_mode
        self.check_mode = False
        self.no_log = no_log
        self.cleanup_files = []
        self._debug = False
        self._diff = False
        self._verbosity = 0
        # May be used to set modifications to the environment for any
        # run_command invocation
        self.run_command_environ_update = {}

        self.aliases = {}
        self._legal_inputs = ['_ansible_check_mode', '_ansible_no_log', '_ansible_debug', '_ansible_diff', '_ansible_verbosity', '_ansible_selinux_special_fs', '_ansible_module_name', '_ansible_version', '_ansible_syslog_facility']

        if add_file_common_args:
            for k, v in FILE_COMMON_ARGUMENTS.items():
                if k not in self.argument_spec:
                    self.argument_spec[k] = v

        self._load_params()
        self._set_fallbacks()

        # append to legal_inputs and then possibly check against them
        try:
            self.aliases = self._handle_aliases()
        except Exception:
            e = get_exception()
            # Use exceptions here because it isn't safe to call fail_json until no_log is processed
            print('\n{"failed": true, "msg": "Module alias error: %s"}' % str(e))
            sys.exit(1)

        # Save parameter values that should never be logged
        self.no_log_values = set()
        # Use the argspec to determine which args are no_log
        for arg_name, arg_opts in self.argument_spec.items():
            if arg_opts.get('no_log', False):
                # Find the value for the no_log'd param
                no_log_object = self.params.get(arg_name, None)
                if no_log_object:
                    self.no_log_values.update(return_values(no_log_object))

        # check the locale as set by the current environment, and reset to
        # a known valid (LANG=C) if it's an invalid/unavailable locale
        self._check_locale()

        self._check_arguments(check_invalid_arguments)

        # check exclusive early
        if not bypass_checks:
            self._check_mutually_exclusive(mutually_exclusive)

        self._set_defaults(pre=True)


        self._CHECK_ARGUMENT_TYPES_DISPATCHER = {
                'str': self._check_type_str,
                'list': self._check_type_list,
                'dict': self._check_type_dict,
                'bool': self._check_type_bool,
                'int': self._check_type_int,
                'float': self._check_type_float,
                'path': self._check_type_path,
                'raw': self._check_type_raw,
                'jsonarg': self._check_type_jsonarg,
            }
        if not bypass_checks:
            self._check_required_arguments()
            self._check_argument_types()
            self._check_argument_values()
            self._check_required_together(required_together)
            self._check_required_one_of(required_one_of)
            self._check_required_if(required_if)

        self._set_defaults(pre=False)

        if not self.no_log and self._verbosity >= 3:
            self._log_invocation()

        # finally, make sure we're in a sane working dir
        self._set_cwd()

    def load_file_common_arguments(self, params):
        '''
        many modules deal with files, this encapsulates common
        options that the file module accepts such that it is directly
        available to all modules and they can share code.
        '''

        path = params.get('path', params.get('dest', None))
        if path is None:
            return {}
        else:
            path = os.path.expanduser(path)

        # if the path is a symlink, and we're following links, get
        # the target of the link instead for testing
        if params.get('follow', False) and os.path.islink(path):
            path = os.path.realpath(path)

        mode   = params.get('mode', None)
        owner  = params.get('owner', None)
        group  = params.get('group', None)

        # selinux related options
        seuser    = params.get('seuser', None)
        serole    = params.get('serole', None)
        setype    = params.get('setype', None)
        selevel   = params.get('selevel', None)
        secontext = [seuser, serole, setype]

        if self.selinux_mls_enabled():
            secontext.append(selevel)

        default_secontext = self.selinux_default_context(path)
        for i in range(len(default_secontext)):
            if i is not None and secontext[i] == '_default':
                secontext[i] = default_secontext[i]

        return dict(
            path=path, mode=mode, owner=owner, group=group,
            seuser=seuser, serole=serole, setype=setype,
            selevel=selevel, secontext=secontext,
        )


    # Detect whether using selinux that is MLS-aware.
    # While this means you can set the level/range with
    # selinux.lsetfilecon(), it may or may not mean that you
    # will get the selevel as part of the context returned
    # by selinux.lgetfilecon().

    def selinux_mls_enabled(self):
        if not HAVE_SELINUX:
            return False
        if selinux.is_selinux_mls_enabled() == 1:
            return True
        else:
            return False

    def selinux_enabled(self):
        if not HAVE_SELINUX:
            seenabled = self.get_bin_path('selinuxenabled')
            if seenabled is not None:
                (rc,out,err) = self.run_command(seenabled)
                if rc == 0:
                    self.fail_json(msg="Aborting, target uses selinux but python bindings (libselinux-python) aren't installed!")
            return False
        if selinux.is_selinux_enabled() == 1:
            return True
        else:
            return False

    # Determine whether we need a placeholder for selevel/mls
    def selinux_initial_context(self):
        context = [None, None, None]
        if self.selinux_mls_enabled():
            context.append(None)
        return context

    def _to_filesystem_str(self, path):
        '''Returns filesystem path as a str, if it wasn't already.

        Used in selinux interactions because it cannot accept unicode
        instances, and specifying complex args in a playbook leaves
        you with unicode instances.  This method currently assumes
        that your filesystem encoding is UTF-8.

        '''
        if isinstance(path, unicode):
            path = path.encode("utf-8")
        return path

    # If selinux fails to find a default, return an array of None
    def selinux_default_context(self, path, mode=0):
        context = self.selinux_initial_context()
        if not HAVE_SELINUX or not self.selinux_enabled():
            return context
        try:
            ret = selinux.matchpathcon(self._to_filesystem_str(path), mode)
        except OSError:
            return context
        if ret[0] == -1:
            return context
        # Limit split to 4 because the selevel, the last in the list,
        # may contain ':' characters
        context = ret[1].split(':', 3)
        return context

    def selinux_context(self, path):
        context = self.selinux_initial_context()
        if not HAVE_SELINUX or not self.selinux_enabled():
            return context
        try:
            ret = selinux.lgetfilecon_raw(self._to_filesystem_str(path))
        except OSError:
            e = get_exception()
            if e.errno == errno.ENOENT:
                self.fail_json(path=path, msg='path %s does not exist' % path)
            else:
                self.fail_json(path=path, msg='failed to retrieve selinux context')
        if ret[0] == -1:
            return context
        # Limit split to 4 because the selevel, the last in the list,
        # may contain ':' characters
        context = ret[1].split(':', 3)
        return context

    def user_and_group(self, filename):
        filename = os.path.expanduser(filename)
        st = os.lstat(filename)
        uid = st.st_uid
        gid = st.st_gid
        return (uid, gid)

    def find_mount_point(self, path):
        path = os.path.realpath(os.path.expanduser(os.path.expandvars(path)))
        while not os.path.ismount(path):
            path = os.path.dirname(path)
        return path

    def is_special_selinux_path(self, path):
        """
        Returns a tuple containing (True, selinux_context) if the given path is on a
        NFS or other 'special' fs  mount point, otherwise the return will be (False, None).
        """
        try:
            f = open('/proc/mounts', 'r')
            mount_data = f.readlines()
            f.close()
        except:
            return (False, None)
        path_mount_point = self.find_mount_point(path)
        for line in mount_data:
            (device, mount_point, fstype, options, rest) = line.split(' ', 4)

            if path_mount_point == mount_point:
                for fs in self._selinux_special_fs:
                    if fs in fstype:
                        special_context = self.selinux_context(path_mount_point)
                        return (True, special_context)

        return (False, None)

    def set_default_selinux_context(self, path, changed):
        if not HAVE_SELINUX or not self.selinux_enabled():
            return changed
        context = self.selinux_default_context(path)
        return self.set_context_if_different(path, context, False)

    def set_context_if_different(self, path, context, changed, diff=None):

        if not HAVE_SELINUX or not self.selinux_enabled():
            return changed
        cur_context = self.selinux_context(path)
        new_context = list(cur_context)
        # Iterate over the current context instead of the
        # argument context, which may have selevel.

        (is_special_se, sp_context) = self.is_special_selinux_path(path)
        if is_special_se:
            new_context = sp_context
        else:
            for i in range(len(cur_context)):
                if len(context) > i:
                    if context[i] is not None and context[i] != cur_context[i]:
                        new_context[i] = context[i]
                    elif context[i] is None:
                        new_context[i] = cur_context[i]

        if cur_context != new_context:
            if diff is not None:
                if 'before' not in diff:
                    diff['before'] = {}
                diff['before']['secontext'] = cur_context
                if 'after' not in diff:
                    diff['after'] = {}
                diff['after']['secontext'] = new_context

            try:
                if self.check_mode:
                    return True
                rc = selinux.lsetfilecon(self._to_filesystem_str(path),
                                         str(':'.join(new_context)))
            except OSError:
                e = get_exception()
                self.fail_json(path=path, msg='invalid selinux context: %s' % str(e), new_context=new_context, cur_context=cur_context, input_was=context)
            if rc != 0:
                self.fail_json(path=path, msg='set selinux context failed')
            changed = True
        return changed

    def set_owner_if_different(self, path, owner, changed, diff=None):
        path = os.path.expanduser(path)
        if owner is None:
            return changed
        orig_uid, orig_gid = self.user_and_group(path)
        try:
            uid = int(owner)
        except ValueError:
            try:
                uid = pwd.getpwnam(owner).pw_uid
            except KeyError:
                self.fail_json(path=path, msg='chown failed: failed to look up user %s' % owner)
        if orig_uid != uid:

            if diff is not None:
                if 'before' not in diff:
                    diff['before'] = {}
                diff['before']['owner'] = orig_uid
                if 'after' not in diff:
                    diff['after'] = {}
                diff['after']['owner'] = uid

            if self.check_mode:
                return True
            try:
                os.lchown(path, uid, -1)
            except OSError:
                self.fail_json(path=path, msg='chown failed')
            changed = True
        return changed

    def set_group_if_different(self, path, group, changed, diff=None):
        path = os.path.expanduser(path)
        if group is None:
            return changed
        orig_uid, orig_gid = self.user_and_group(path)
        try:
            gid = int(group)
        except ValueError:
            try:
                gid = grp.getgrnam(group).gr_gid
            except KeyError:
                self.fail_json(path=path, msg='chgrp failed: failed to look up group %s' % group)
        if orig_gid != gid:

            if diff is not None:
                if 'before' not in diff:
                    diff['before'] = {}
                diff['before']['group'] = orig_gid
                if 'after' not in diff:
                    diff['after'] = {}
                diff['after']['group'] = gid

            if self.check_mode:
                return True
            try:
                os.lchown(path, -1, gid)
            except OSError:
                self.fail_json(path=path, msg='chgrp failed')
            changed = True
        return changed

    def set_mode_if_different(self, path, mode, changed, diff=None):
        path = os.path.expanduser(path)
        path_stat = os.lstat(path)

        if mode is None:
            return changed

        if not isinstance(mode, int):
            try:
                mode = int(mode, 8)
            except Exception:
                try:
                    mode = self._symbolic_mode_to_octal(path_stat, mode)
                except Exception:
                    e = get_exception()
                    self.fail_json(path=path,
                                   msg="mode must be in octal or symbolic form",
                                   details=str(e))

                if mode != stat.S_IMODE(mode):
                    # prevent mode from having extra info orbeing invalid long number
                    self.fail_json(path=path, msg="Invalid mode supplied, only permission info is allowed", details=mode)

        prev_mode = stat.S_IMODE(path_stat.st_mode)

        if prev_mode != mode:

            if diff is not None:
                if 'before' not in diff:
                    diff['before'] = {}
                diff['before']['mode'] = oct(prev_mode)
                if 'after' not in diff:
                    diff['after'] = {}
                diff['after']['mode'] = oct(mode)

            if self.check_mode:
                return True
            # FIXME: comparison against string above will cause this to be executed
            # every time
            try:
                if hasattr(os, 'lchmod'):
                    os.lchmod(path, mode)
                else:
                    if not os.path.islink(path):
                        os.chmod(path, mode)
                    else:
                        # Attempt to set the perms of the symlink but be
                        # careful not to change the perms of the underlying
                        # file while trying
                        underlying_stat = os.stat(path)
                        os.chmod(path, mode)
                        new_underlying_stat = os.stat(path)
                        if underlying_stat.st_mode != new_underlying_stat.st_mode:
                            os.chmod(path, stat.S_IMODE(underlying_stat.st_mode))
            except OSError:
                e = get_exception()
                if os.path.islink(path) and e.errno == errno.EPERM:  # Can't set mode on symbolic links
                    pass
                elif e.errno in (errno.ENOENT, errno.ELOOP): # Can't set mode on broken symbolic links
                    pass
                else:
                    raise e
            except Exception:
                e = get_exception()
                self.fail_json(path=path, msg='chmod failed', details=str(e))

            path_stat = os.lstat(path)
            new_mode = stat.S_IMODE(path_stat.st_mode)

            if new_mode != prev_mode:
                changed = True
        return changed

    def _symbolic_mode_to_octal(self, path_stat, symbolic_mode):
        new_mode = stat.S_IMODE(path_stat.st_mode)

        mode_re = re.compile(r'^(?P<users>[ugoa]+)(?P<operator>[-+=])(?P<perms>[rwxXst-]*|[ugo])$')
        for mode in symbolic_mode.split(','):
            match = mode_re.match(mode)
            if match:
                users = match.group('users')
                operator = match.group('operator')
                perms = match.group('perms')

                if users == 'a':
                    users = 'ugo'

                for user in users:
                    mode_to_apply = self._get_octal_mode_from_symbolic_perms(path_stat, user, perms)
                    new_mode = self._apply_operation_to_mode(user, operator, mode_to_apply, new_mode)
            else:
                raise ValueError("bad symbolic permission for mode: %s" % mode)
        return new_mode

    def _apply_operation_to_mode(self, user, operator, mode_to_apply, current_mode):
        if operator  ==  '=':
            if user == 'u': mask = stat.S_IRWXU | stat.S_ISUID
            elif user == 'g': mask = stat.S_IRWXG | stat.S_ISGID
            elif user == 'o': mask = stat.S_IRWXO | stat.S_ISVTX

            # mask out u, g, or o permissions from current_mode and apply new permissions
            inverse_mask = mask ^ PERM_BITS
            new_mode = (current_mode & inverse_mask) | mode_to_apply
        elif operator == '+':
            new_mode = current_mode | mode_to_apply
        elif operator == '-':
            new_mode = current_mode - (current_mode & mode_to_apply)
        return new_mode

    def _get_octal_mode_from_symbolic_perms(self, path_stat, user, perms):
        prev_mode = stat.S_IMODE(path_stat.st_mode)

        is_directory = stat.S_ISDIR(path_stat.st_mode)
        has_x_permissions = (prev_mode & EXEC_PERM_BITS) > 0
        apply_X_permission = is_directory or has_x_permissions

        # Permission bits constants documented at:
        # http://docs.python.org/2/library/stat.html#stat.S_ISUID
        if apply_X_permission:
            X_perms = {
                'u': {'X': stat.S_IXUSR},
                'g': {'X': stat.S_IXGRP},
                'o': {'X': stat.S_IXOTH}
            }
        else:
            X_perms = {
                'u': {'X': 0},
                'g': {'X': 0},
                'o': {'X': 0}
            }

        user_perms_to_modes = {
            'u': {
                'r': stat.S_IRUSR,
                'w': stat.S_IWUSR,
                'x': stat.S_IXUSR,
                's': stat.S_ISUID,
                't': 0,
                'u': prev_mode & stat.S_IRWXU,
                'g': (prev_mode & stat.S_IRWXG) << 3,
                'o': (prev_mode & stat.S_IRWXO) << 6 },
            'g': {
                'r': stat.S_IRGRP,
                'w': stat.S_IWGRP,
                'x': stat.S_IXGRP,
                's': stat.S_ISGID,
                't': 0,
                'u': (prev_mode & stat.S_IRWXU) >> 3,
                'g': prev_mode & stat.S_IRWXG,
                'o': (prev_mode & stat.S_IRWXO) << 3 },
            'o': {
                'r': stat.S_IROTH,
                'w': stat.S_IWOTH,
                'x': stat.S_IXOTH,
                's': 0,
                't': stat.S_ISVTX,
                'u': (prev_mode & stat.S_IRWXU) >> 6,
                'g': (prev_mode & stat.S_IRWXG) >> 3,
                'o': prev_mode & stat.S_IRWXO }
        }

        # Insert X_perms into user_perms_to_modes
        for key, value in X_perms.items():
            user_perms_to_modes[key].update(value)

        or_reduce = lambda mode, perm: mode | user_perms_to_modes[user][perm]
        return reduce(or_reduce, perms, 0)

    def set_fs_attributes_if_different(self, file_args, changed, diff=None):
        # set modes owners and context as needed
        changed = self.set_context_if_different(
            file_args['path'], file_args['secontext'], changed, diff
        )
        changed = self.set_owner_if_different(
            file_args['path'], file_args['owner'], changed, diff
        )
        changed = self.set_group_if_different(
            file_args['path'], file_args['group'], changed, diff
        )
        changed = self.set_mode_if_different(
            file_args['path'], file_args['mode'], changed, diff
        )
        return changed

    def set_directory_attributes_if_different(self, file_args, changed, diff=None):
        return self.set_fs_attributes_if_different(file_args, changed, diff)

    def set_file_attributes_if_different(self, file_args, changed, diff=None):
        return self.set_fs_attributes_if_different(file_args, changed, diff)

    def add_path_info(self, kwargs):
        '''
        for results that are files, supplement the info about the file
        in the return path with stats about the file path.
        '''

        path = kwargs.get('path', kwargs.get('dest', None))
        if path is None:
            return kwargs
        if os.path.exists(path):
            (uid, gid) = self.user_and_group(path)
            kwargs['uid'] = uid
            kwargs['gid'] = gid
            try:
                user = pwd.getpwuid(uid)[0]
            except KeyError:
                user = str(uid)
            try:
                group = grp.getgrgid(gid)[0]
            except KeyError:
                group = str(gid)
            kwargs['owner'] = user
            kwargs['group'] = group
            st = os.lstat(path)
            kwargs['mode']  = oct(stat.S_IMODE(st[stat.ST_MODE]))
            # secontext not yet supported
            if os.path.islink(path):
                kwargs['state'] = 'link'
            elif os.path.isdir(path):
                kwargs['state'] = 'directory'
            elif os.stat(path).st_nlink > 1:
                kwargs['state'] = 'hard'
            else:
                kwargs['state'] = 'file'
            if HAVE_SELINUX and self.selinux_enabled():
                kwargs['secontext'] = ':'.join(self.selinux_context(path))
            kwargs['size'] = st[stat.ST_SIZE]
        else:
            kwargs['state'] = 'absent'
        return kwargs

    def _check_locale(self):
        '''
        Uses the locale module to test the currently set locale
        (per the LANG and LC_CTYPE environment settings)
        '''
        try:
            # setting the locale to '' uses the default locale
            # as it would be returned by locale.getdefaultlocale()
            locale.setlocale(locale.LC_ALL, '')
        except locale.Error:
            # fallback to the 'C' locale, which may cause unicode
            # issues but is preferable to simply failing because
            # of an unknown locale
            locale.setlocale(locale.LC_ALL, 'C')
            os.environ['LANG'] = 'C'
            os.environ['LC_ALL'] = 'C'
            os.environ['LC_MESSAGES'] = 'C'
        except Exception:
            e = get_exception()
            self.fail_json(msg="An unknown error was encountered while attempting to validate the locale: %s" % e)

    def _handle_aliases(self):
        # this uses exceptions as it happens before we can safely call fail_json
        aliases_results = {} #alias:canon
        for (k,v) in self.argument_spec.items():
            self._legal_inputs.append(k)
            aliases = v.get('aliases', None)
            default = v.get('default', None)
            required = v.get('required', False)
            if default is not None and required:
                # not alias specific but this is a good place to check this
                raise Exception("internal error: required and default are mutually exclusive for %s" % k)
            if aliases is None:
                continue
            if type(aliases) != list:
                raise Exception('internal error: aliases must be a list')
            for alias in aliases:
                self._legal_inputs.append(alias)
                aliases_results[alias] = k
                if alias in self.params:
                    self.params[k] = self.params[alias]

        return aliases_results

    def _check_arguments(self, check_invalid_arguments):
        self._syslog_facility = 'LOG_USER'
        for (k,v) in list(self.params.items()):

            if k == '_ansible_check_mode' and v:
                self.check_mode = True

            elif k == '_ansible_no_log':
                self.no_log = self.boolean(v)

            elif k == '_ansible_debug':
                self._debug = self.boolean(v)

            elif k == '_ansible_diff':
                self._diff = self.boolean(v)

            elif k == '_ansible_verbosity':
                self._verbosity = v

            elif k == '_ansible_selinux_special_fs':
                self._selinux_special_fs = v

            elif k == '_ansible_syslog_facility':
                self._syslog_facility = v

            elif k == '_ansible_version':
                self.ansible_version = v

            elif k == '_ansible_module_name':
                self._name = v

            elif check_invalid_arguments and k not in self._legal_inputs:
                self.fail_json(msg="unsupported parameter for module: %s" % k)

            #clean up internal params:
            if k.startswith('_ansible_'):
                del self.params[k]

        if self.check_mode and not self.supports_check_mode:
                self.exit_json(skipped=True, msg="remote module (%s) does not support check mode" % self._name)

    def _count_terms(self, check):
        count = 0
        for term in check:
            if term in self.params:
                count += 1
        return count

    def _check_mutually_exclusive(self, spec):
        if spec is None:
            return
        for check in spec:
            count = self._count_terms(check)
            if count > 1:
                self.fail_json(msg="parameters are mutually exclusive: %s" % (check,))

    def _check_required_one_of(self, spec):
        if spec is None:
            return
        for check in spec:
            count = self._count_terms(check)
            if count == 0:
                self.fail_json(msg="one of the following is required: %s" % ','.join(check))

    def _check_required_together(self, spec):
        if spec is None:
            return
        for check in spec:
            counts = [ self._count_terms([field]) for field in check ]
            non_zero = [ c for c in counts if c > 0 ]
            if len(non_zero) > 0:
                if 0 in counts:
                    self.fail_json(msg="parameters are required together: %s" % (check,))

    def _check_required_arguments(self):
        ''' ensure all required arguments are present '''
        missing = []
        for (k,v) in self.argument_spec.items():
            required = v.get('required', False)
            if required and k not in self.params:
                missing.append(k)
        if len(missing) > 0:
            self.fail_json(msg="missing required arguments: %s" % ",".join(missing))

    def _check_required_if(self, spec):
        ''' ensure that parameters which conditionally required are present '''
        if spec is None:
            return
        for (key, val, requirements) in spec:
            missing = []
            if key in self.params and self.params[key] == val:
                for check in requirements:
                    count = self._count_terms((check,))
                    if count == 0:
                        missing.append(check)
            if len(missing) > 0:
                self.fail_json(msg="%s is %s but the following are missing: %s" % (key, val, ','.join(missing)))

    def _check_argument_values(self):
        ''' ensure all arguments have the requested values, and there are no stray arguments '''
        for (k,v) in self.argument_spec.items():
            choices = v.get('choices',None)
            if choices is None:
                continue
            if isinstance(choices, SEQUENCETYPE):
                if k in self.params:
                    if self.params[k] not in choices:
                        choices_str=",".join([str(c) for c in choices])
                        msg="value of %s must be one of: %s, got: %s" % (k, choices_str, self.params[k])
                        self.fail_json(msg=msg)
            else:
                self.fail_json(msg="internal error: choices for argument %s are not iterable: %s" % (k, choices))

    def safe_eval(self, str, locals=None, include_exceptions=False):

        # do not allow method calls to modules
        if not isinstance(str, basestring):
            # already templated to a datastructure, perhaps?
            if include_exceptions:
                return (str, None)
            return str
        if re.search(r'\w\.\w+\(', str):
            if include_exceptions:
                return (str, None)
            return str
        # do not allow imports
        if re.search(r'import \w+', str):
            if include_exceptions:
                return (str, None)
            return str
        try:
            result = literal_eval(str)
            if include_exceptions:
                return (result, None)
            else:
                return result
        except Exception:
            e = get_exception()
            if include_exceptions:
                return (str, e)
            return str

    def _check_type_str(self, value):
        if isinstance(value, basestring):
            return value
        # Note: This could throw a unicode error if value's __str__() method
        # returns non-ascii.  Have to port utils.to_bytes() if that happens
        return str(value)

    def _check_type_list(self, value):
        if isinstance(value, list):
            return value

        if isinstance(value, basestring):
            return value.split(",")
        elif isinstance(value, int) or isinstance(value, float):
            return [ str(value) ]

        raise TypeError('%s cannot be converted to a list' % type(value))

    def _check_type_dict(self, value):
        if isinstance(value, dict):
            return value

        if isinstance(value, basestring):
            if value.startswith("{"):
                try:
                    return json.loads(value)
                except:
                    (result, exc) = self.safe_eval(value, dict(), include_exceptions=True)
                    if exc is not None:
                        raise TypeError('unable to evaluate string as dictionary')
                    return result
            elif '=' in value:
                fields = []
                field_buffer = []
                in_quote = False
                in_escape = False
                for c in value.strip():
                    if in_escape:
                        field_buffer.append(c)
                        in_escape = False
                    elif c == '\\':
                        in_escape = True
                    elif not in_quote and c in ('\'', '"'):
                        in_quote = c
                    elif in_quote and in_quote == c:
                        in_quote = False
                    elif not in_quote and c in (',', ' '):
                        field = ''.join(field_buffer)
                        if field:
                            fields.append(field)
                        field_buffer = []
                    else:
                        field_buffer.append(c)

                field = ''.join(field_buffer)
                if field:
                    fields.append(field)
                return dict(x.split("=", 1) for x in fields)
            else:
                raise TypeError("dictionary requested, could not parse JSON or key=value")

        raise TypeError('%s cannot be converted to a dict' % type(value))

    def _check_type_bool(self, value):
        if isinstance(value, bool):
            return value

        if isinstance(value, basestring) or isinstance(value, int):
            return self.boolean(value)

        raise TypeError('%s cannot be converted to a bool' % type(value))

    def _check_type_int(self, value):
        if isinstance(value, int):
            return value

        if isinstance(value, basestring):
            return int(value)

        raise TypeError('%s cannot be converted to an int' % type(value))

    def _check_type_float(self, value):
        if isinstance(value, float):
            return value

        if isinstance(value, basestring):
            return float(value)

        raise TypeError('%s cannot be converted to a float' % type(value))

    def _check_type_path(self, value):
        value = self._check_type_str(value)
        return os.path.expanduser(os.path.expandvars(value))

    def _check_type_jsonarg(self, value):
        # Return a jsonified string.  Sometimes the controller turns a json
        # string into a dict/list so transform it back into json here
        if isinstance(value, (unicode, bytes)):
            return value.strip()
        else:
            if isinstance(value (list, tuple, dict)):
                return json.dumps(value)
        raise TypeError('%s cannot be converted to a json string' % type(value))

    def _check_type_raw(self, value):
        return value


    def _check_argument_types(self):
        ''' ensure all arguments have the requested type '''
        for (k, v) in self.argument_spec.items():
            wanted = v.get('type', None)
            if k not in self.params:
                continue
            if wanted is None:
                # Mostly we want to default to str.
                # For values set to None explicitly, return None instead as
                # that allows a user to unset a parameter
                if self.params[k] is None:
                    continue
                wanted = 'str'

            value = self.params[k]
            if value is None:
                continue

            try:
                type_checker = self._CHECK_ARGUMENT_TYPES_DISPATCHER[wanted]
            except KeyError:
                self.fail_json(msg="implementation error: unknown type %s requested for %s" % (wanted, k))
            try:
                self.params[k] = type_checker(value)
            except (TypeError, ValueError):
                self.fail_json(msg="argument %s is of type %s and we were unable to convert to %s" % (k, type(value), wanted))

    def _set_defaults(self, pre=True):
        for (k,v) in self.argument_spec.items():
            default = v.get('default', None)
            if pre == True:
                # this prevents setting defaults on required items
                if default is not None and k not in self.params:
                    self.params[k] = default
            else:
                # make sure things without a default still get set None
                if k not in self.params:
                    self.params[k] = default

    def _set_fallbacks(self):
        for k,v in self.argument_spec.items():
            fallback = v.get('fallback', (None,))
            fallback_strategy = fallback[0]
            fallback_args = []
            fallback_kwargs = {}
            if k not in self.params and fallback_strategy is not None:
                for item in fallback[1:]:
                    if isinstance(item, dict):
                        fallback_kwargs = item
                    else:
                        fallback_args = item
                try:
                    self.params[k] = fallback_strategy(*fallback_args, **fallback_kwargs)
                except AnsibleFallbackNotFound:
                    continue

    def _load_params(self):
        ''' read the input and set the params attribute.

        This method is for backwards compatibility.  The guts of the function
        were moved out in 2.1 so that custom modules could read the parameters.
        '''
        # debug overrides to read args from file or cmdline
        self.params = _load_params()

    def _log_to_syslog(self, msg):
        if HAS_SYSLOG:
            module = 'ansible-%s' % self._name
            facility = getattr(syslog, self._syslog_facility, syslog.LOG_USER)
            syslog.openlog(str(module), 0, facility)
            syslog.syslog(syslog.LOG_INFO, msg)

    def debug(self, msg):
        if self._debug:
            self.log(msg)

    def log(self, msg, log_args=None):

        if not self.no_log:

            if log_args is None:
                log_args = dict()

            module = 'ansible-%s' % self._name
            if isinstance(module, bytes):
                module = module.decode('utf-8', 'replace')

            # 6655 - allow for accented characters
            if not isinstance(msg, (bytes, unicode)):
                raise TypeError("msg should be a string (got %s)" % type(msg))

            # We want journal to always take text type
            # syslog takes bytes on py2, text type on py3
            if isinstance(msg, bytes):
                journal_msg = remove_values(msg.decode('utf-8', 'replace'), self.no_log_values)
            else:
                # TODO: surrogateescape is a danger here on Py3
                journal_msg = remove_values(msg, self.no_log_values)

            if PY3:
                syslog_msg = journal_msg
            else:
                syslog_msg = journal_msg.encode('utf-8', 'replace')

            if has_journal:
                journal_args = [("MODULE", os.path.basename(__file__))]
                for arg in log_args:
                    journal_args.append((arg.upper(), str(log_args[arg])))
                try:
                    journal.send(u"%s %s" % (module, journal_msg), **dict(journal_args))
                except IOError:
                    # fall back to syslog since logging to journal failed
                    self._log_to_syslog(syslog_msg)
            else:
                self._log_to_syslog(syslog_msg)

    def _log_invocation(self):
        ''' log that ansible ran the module '''
        # TODO: generalize a separate log function and make log_invocation use it
        # Sanitize possible password argument when logging.
        log_args = dict()
        passwd_keys = ['password', 'login_password']

        for param in self.params:
            canon  = self.aliases.get(param, param)
            arg_opts = self.argument_spec.get(canon, {})
            no_log = arg_opts.get('no_log', False)

            if self.boolean(no_log):
                log_args[param] = 'NOT_LOGGING_PARAMETER'
            elif param in passwd_keys:
                log_args[param] = 'NOT_LOGGING_PASSWORD'
            else:
                param_val = self.params[param]
                if not isinstance(param_val, basestring):
                    param_val = str(param_val)
                elif isinstance(param_val, unicode):
                    param_val = param_val.encode('utf-8')
                log_args[param] = heuristic_log_sanitize(param_val, self.no_log_values)

        msg = []
        for arg in log_args:
            arg_val = log_args[arg]
            if not isinstance(arg_val, basestring):
                arg_val = str(arg_val)
            elif isinstance(arg_val, unicode):
                arg_val = arg_val.encode('utf-8')
            msg.append('%s=%s' % (arg, arg_val))
        if msg:
            msg = 'Invoked with %s' % ' '.join(msg)
        else:
            msg = 'Invoked'

        self.log(msg, log_args=log_args)


    def _set_cwd(self):
        try:
            cwd = os.getcwd()
            if not os.access(cwd, os.F_OK|os.R_OK):
                raise
            return cwd
        except:
            # we don't have access to the cwd, probably because of sudo.
            # Try and move to a neutral location to prevent errors
            for cwd in [os.path.expandvars('$HOME'), tempfile.gettempdir()]:
                try:
                    if os.access(cwd, os.F_OK|os.R_OK):
                        os.chdir(cwd)
                        return cwd
                except:
                    pass
        # we won't error here, as it may *not* be a problem,
        # and we don't want to break modules unnecessarily
        return None

    def get_bin_path(self, arg, required=False, opt_dirs=[]):
        '''
        find system executable in PATH.
        Optional arguments:
           - required:  if executable is not found and required is true, fail_json
           - opt_dirs:  optional list of directories to search in addition to PATH
        if found return full path; otherwise return None
        '''
        sbin_paths = ['/sbin', '/usr/sbin', '/usr/local/sbin']
        paths = []
        for d in opt_dirs:
            if d is not None and os.path.exists(d):
                paths.append(d)
        paths += os.environ.get('PATH', '').split(os.pathsep)
        bin_path = None
        # mangle PATH to include /sbin dirs
        for p in sbin_paths:
            if p not in paths and os.path.exists(p):
                paths.append(p)
        for d in paths:
            if not d:
                continue
            path = os.path.join(d, arg)
            if os.path.exists(path) and is_executable(path):
                bin_path = path
                break
        if required and bin_path is None:
            self.fail_json(msg='Failed to find required executable %s' % arg)
        return bin_path

    def boolean(self, arg):
        ''' return a bool for the arg '''
        if arg is None or type(arg) == bool:
            return arg
        if isinstance(arg, basestring):
            arg = arg.lower()
        if arg in BOOLEANS_TRUE:
            return True
        elif arg in BOOLEANS_FALSE:
            return False
        else:
            self.fail_json(msg='Boolean %s not in either boolean list' % arg)

    def jsonify(self, data):
        for encoding in ("utf-8", "latin-1"):
            try:
                return json.dumps(data, encoding=encoding)
            # Old systems using old simplejson module does not support encoding keyword.
            except TypeError:
                try:
                    new_data = json_dict_bytes_to_unicode(data, encoding=encoding)
                except UnicodeDecodeError:
                    continue
                return json.dumps(new_data)
            except UnicodeDecodeError:
                continue
        self.fail_json(msg='Invalid unicode encoding encountered')

    def from_json(self, data):
        return json.loads(data)

    def add_cleanup_file(self, path):
        if path not in self.cleanup_files:
            self.cleanup_files.append(path)

    def do_cleanup_files(self):
        for path in self.cleanup_files:
            self.cleanup(path)

    def exit_json(self, **kwargs):
        ''' return from the module, without error '''
        self.add_path_info(kwargs)
        if not 'changed' in kwargs:
            kwargs['changed'] = False
        if 'invocation' not in kwargs:
            kwargs['invocation'] = {'module_args': self.params}
        kwargs = remove_values(kwargs, self.no_log_values)
        self.do_cleanup_files()
        print('\n%s' % self.jsonify(kwargs))
        sys.exit(0)

    def fail_json(self, **kwargs):
        ''' return from the module, with an error message '''
        self.add_path_info(kwargs)
        assert 'msg' in kwargs, "implementation error -- msg to explain the error is required"
        kwargs['failed'] = True
        if 'invocation' not in kwargs:
            kwargs['invocation'] = {'module_args': self.params}
        kwargs = remove_values(kwargs, self.no_log_values)
        self.do_cleanup_files()
        print('\n%s' % self.jsonify(kwargs))
        sys.exit(1)

    def fail_on_missing_params(self, required_params=None):
        ''' This is for checking for required params when we can not check via argspec because we
        need more information than is simply given in the argspec.
        '''
        if not required_params:
            return
        missing_params = []
        for required_param in required_params:
            if not self.params.get(required_param):
                missing_params.append(required_param)
        if missing_params:
            self.fail_json(msg="missing required arguments: %s" % ','.join(missing_params))

    def digest_from_file(self, filename, algorithm):
        ''' Return hex digest of local file for a digest_method specified by name, or None if file is not present. '''
        if not os.path.exists(filename):
            return None
        if os.path.isdir(filename):
            self.fail_json(msg="attempted to take checksum of directory: %s" % filename)

        # preserve old behaviour where the third parameter was a hash algorithm object
        if hasattr(algorithm, 'hexdigest'):
            digest_method = algorithm
        else:
            try:
                digest_method = AVAILABLE_HASH_ALGORITHMS[algorithm]()
            except KeyError:
                self.fail_json(msg="Could not hash file '%s' with algorithm '%s'. Available algorithms: %s" %
                                   (filename, algorithm, ', '.join(AVAILABLE_HASH_ALGORITHMS)))

        blocksize = 64 * 1024
        infile = open(filename, 'rb')
        block = infile.read(blocksize)
        while block:
            digest_method.update(block)
            block = infile.read(blocksize)
        infile.close()
        return digest_method.hexdigest()

    def md5(self, filename):
        ''' Return MD5 hex digest of local file using digest_from_file().

        Do not use this function unless you have no other choice for:
            1) Optional backwards compatibility
            2) Compatibility with a third party protocol

        This function will not work on systems complying with FIPS-140-2.

        Most uses of this function can use the module.sha1 function instead.
        '''
        if 'md5' not in AVAILABLE_HASH_ALGORITHMS:
            raise ValueError('MD5 not available.  Possibly running in FIPS mode')
        return self.digest_from_file(filename, 'md5')

    def sha1(self, filename):
        ''' Return SHA1 hex digest of local file using digest_from_file(). '''
        return self.digest_from_file(filename, 'sha1')

    def sha256(self, filename):
        ''' Return SHA-256 hex digest of local file using digest_from_file(). '''
        return self.digest_from_file(filename, 'sha256')

    def backup_local(self, fn):
        '''make a date-marked backup of the specified file, return True or False on success or failure'''

        backupdest = ''
        if os.path.exists(fn):
            # backups named basename-YYYY-MM-DD@HH:MM:SS~
            ext = time.strftime("%Y-%m-%d@%H:%M:%S~", time.localtime(time.time()))
            backupdest = '%s.%s' % (fn, ext)

            try:
                shutil.copy2(fn, backupdest)
            except (shutil.Error, IOError):
                e = get_exception()
                self.fail_json(msg='Could not make backup of %s to %s: %s' % (fn, backupdest, e))

        return backupdest

    def cleanup(self, tmpfile):
        if os.path.exists(tmpfile):
            try:
                os.unlink(tmpfile)
            except OSError:
                e = get_exception()
                sys.stderr.write("could not cleanup %s: %s" % (tmpfile, e))

    def atomic_move(self, src, dest, unsafe_writes=False):
        '''atomically move src to dest, copying attributes from dest, returns true on success
        it uses os.rename to ensure this as it is an atomic operation, rest of the function is
        to work around limitations, corner cases and ensure selinux context is saved if possible'''
        context = None
        dest_stat = None
        if os.path.exists(dest):
            try:
                dest_stat = os.stat(dest)
                os.chmod(src, dest_stat.st_mode & PERM_BITS)
                os.chown(src, dest_stat.st_uid, dest_stat.st_gid)
            except OSError:
                e = get_exception()
                if e.errno != errno.EPERM:
                    raise
            if self.selinux_enabled():
                context = self.selinux_context(dest)
        else:
            if self.selinux_enabled():
                context = self.selinux_default_context(dest)

        creating = not os.path.exists(dest)

        try:
            login_name = os.getlogin()
        except OSError:
            # not having a tty can cause the above to fail, so
            # just get the LOGNAME environment variable instead
            login_name = os.environ.get('LOGNAME', None)

        # if the original login_name doesn't match the currently
        # logged-in user, or if the SUDO_USER environment variable
        # is set, then this user has switched their credentials
        switched_user = login_name and login_name != pwd.getpwuid(os.getuid())[0] or os.environ.get('SUDO_USER')

        try:
            # Optimistically try a rename, solves some corner cases and can avoid useless work, throws exception if not atomic.
            os.rename(src, dest)
        except (IOError, OSError):
            e = get_exception()
            if e.errno not in [errno.EPERM, errno.EXDEV, errno.EACCES, errno.ETXTBSY]:
                # only try workarounds for errno 18 (cross device), 1 (not permitted),  13 (permission denied)
                # and 26 (text file busy) which happens on vagrant synced folders and other 'exotic' non posix file systems
                self.fail_json(msg='Could not replace file: %s to %s: %s' % (src, dest, e))
            else:
                dest_dir = os.path.dirname(dest)
                dest_file = os.path.basename(dest)
                try:
                    tmp_dest = tempfile.NamedTemporaryFile(
                        prefix=".ansible_tmp", dir=dest_dir, suffix=dest_file)
                except (OSError, IOError):
                    e = get_exception()
                    self.fail_json(msg='The destination directory (%s) is not writable by the current user. Error was: %s' % (dest_dir, e))

                try: # leaves tmp file behind when sudo and  not root
                    if switched_user and os.getuid() != 0:
                        # cleanup will happen by 'rm' of tempdir
                        # copy2 will preserve some metadata
                        shutil.copy2(src, tmp_dest.name)
                    else:
                        shutil.move(src, tmp_dest.name)
                    if self.selinux_enabled():
                        self.set_context_if_different(
                            tmp_dest.name, context, False)
                    try:
                        tmp_stat = os.stat(tmp_dest.name)
                        if dest_stat and (tmp_stat.st_uid != dest_stat.st_uid or tmp_stat.st_gid != dest_stat.st_gid):
                            os.chown(tmp_dest.name, dest_stat.st_uid, dest_stat.st_gid)
                    except OSError:
                        e = get_exception()
                        if e.errno != errno.EPERM:
                            raise
                    os.rename(tmp_dest.name, dest)
                except (shutil.Error, OSError, IOError):
                    e = get_exception()
                    # sadly there are some situations where we cannot ensure atomicity, but only if
                    # the user insists and we get the appropriate error we update the file unsafely
                    if unsafe_writes and e.errno == errno.EBUSY:
                        #TODO: issue warning that this is an unsafe operation, but doing it cause user insists
                        try:
                            try:
                                out_dest = open(dest, 'wb')
                                in_src = open(src, 'rb')
                                shutil.copyfileobj(in_src, out_dest)
                            finally: # assuring closed files in 2.4 compatible way
                                if out_dest:
                                    out_dest.close()
                                if in_src:
                                    in_src.close()
                        except (shutil.Error, OSError, IOError):
                            e = get_exception()
                            self.fail_json(msg='Could not write data to file (%s) from (%s): %s' % (dest, src, e))

                    else:
                        self.fail_json(msg='Could not replace file: %s to %s: %s' % (src, dest, e))

                    self.cleanup(tmp_dest.name)

        if creating:
            # make sure the file has the correct permissions
            # based on the current value of umask
            umask = os.umask(0)
            os.umask(umask)
            os.chmod(dest, DEFAULT_PERM & ~umask)
            if switched_user:
                os.chown(dest, os.getuid(), os.getgid())

        if self.selinux_enabled():
            # rename might not preserve context
            self.set_context_if_different(dest, context, False)

    def run_command(self, args, check_rc=False, close_fds=True, executable=None, data=None, binary_data=False, path_prefix=None, cwd=None, use_unsafe_shell=False, prompt_regex=None, environ_update=None):
        '''
        Execute a command, returns rc, stdout, and stderr.

        :arg args: is the command to run
            * If args is a list, the command will be run with shell=False.
            * If args is a string and use_unsafe_shell=False it will split args to a list and run with shell=False
            * If args is a string and use_unsafe_shell=True it runs with shell=True.
        :kw check_rc: Whether to call fail_json in case of non zero RC.
            Default False
        :kw close_fds: See documentation for subprocess.Popen(). Default True
        :kw executable: See documentation for subprocess.Popen(). Default None
        :kw data: If given, information to write to the stdin of the command
        :kw binary_data: If False, append a newline to the data.  Default False
        :kw path_prefix: If given, additional path to find the command in.
            This adds to the PATH environment vairable so helper commands in
            the same directory can also be found
        :kw cwd: iIf given, working directory to run the command inside
        :kw use_unsafe_shell: See `args` parameter.  Default False
        :kw prompt_regex: Regex string (not a compiled regex) which can be
            used to detect prompts in the stdout which would otherwise cause
            the execution to hang (especially if no input data is specified)
        :kwarg environ_update: dictionary to *update* os.environ with
        '''

        shell = False
        if isinstance(args, list):
            if use_unsafe_shell:
                args = " ".join([pipes.quote(x) for x in args])
                shell = True
        elif isinstance(args, basestring) and use_unsafe_shell:
            shell = True
        elif isinstance(args, string_types):
            # On python2.6 and below, shlex has problems with text type
            # On python3, shlex needs a text type.
            if PY2 and isinstance(args, text_type):
                args = args.encode('utf-8')
            elif PY3 and isinstance(args, binary_type):
                args = args.decode('utf-8', errors='surrogateescape')
            args = shlex.split(args)
        else:
            msg = "Argument 'args' to run_command must be list or string"
            self.fail_json(rc=257, cmd=args, msg=msg)

        prompt_re = None
        if prompt_regex:
            if isinstance(prompt_regex, text_type):
                if PY3:
                    prompt_regex = prompt_regex.encode('utf-8', errors='surrogateescape')
                elif PY2:
                    prompt_regex = prompt_regex.encode('utf-8')
            try:
                prompt_re = re.compile(prompt_regex, re.MULTILINE)
            except re.error:
                self.fail_json(msg="invalid prompt regular expression given to run_command")

        # expand things like $HOME and ~
        if not shell:
            args = [ os.path.expandvars(os.path.expanduser(x)) for x in args if x is not None ]

        rc = 0
        msg = None
        st_in = None

        # Manipulate the environ we'll send to the new process
        old_env_vals = {}
        # We can set this from both an attribute and per call
        for key, val in self.run_command_environ_update.items():
            old_env_vals[key] = os.environ.get(key, None)
            os.environ[key] = val
        if environ_update:
            for key, val in environ_update.items():
                old_env_vals[key] = os.environ.get(key, None)
                os.environ[key] = val
        if path_prefix:
            old_env_vals['PATH'] = os.environ['PATH']
            os.environ['PATH'] = "%s:%s" % (path_prefix, os.environ['PATH'])

        # If using test-module and explode, the remote lib path will resemble ...
        #   /tmp/test_module_scratch/debug_dir/ansible/module_utils/basic.py
        # If using ansible or ansible-playbook with a remote system ...
        #   /tmp/ansible_vmweLQ/ansible_modlib.zip/ansible/module_utils/basic.py

        # Clean out python paths set by ansiballz
        if 'PYTHONPATH' in os.environ:
            pypaths = os.environ['PYTHONPATH'].split(':')
            pypaths = [x for x in pypaths \
                        if not x.endswith('/ansible_modlib.zip') \
                        and not x.endswith('/debug_dir')]
            os.environ['PYTHONPATH'] = ':'.join(pypaths)
            if not os.environ['PYTHONPATH']:
                del os.environ['PYTHONPATH']

        # create a printable version of the command for use
        # in reporting later, which strips out things like
        # passwords from the args list
        to_clean_args = args
        if PY2:
            if isinstance(args, text_type):
                to_clean_args = args.encode('utf-8')
        else:
            if isinstance(args, binary_type):
                to_clean_args = args.decode('utf-8', errors='replace')
        if isinstance(args, (text_type, binary_type)):
            to_clean_args = shlex.split(to_clean_args)

        clean_args = []
        is_passwd = False
        for arg in to_clean_args:
            if is_passwd:
                is_passwd = False
                clean_args.append('********')
                continue
            if PASSWD_ARG_RE.match(arg):
                sep_idx = arg.find('=')
                if sep_idx > -1:
                    clean_args.append('%s=********' % arg[:sep_idx])
                    continue
                else:
                    is_passwd = True
            arg = heuristic_log_sanitize(arg, self.no_log_values)
            clean_args.append(arg)
        clean_args = ' '.join(pipes.quote(arg) for arg in clean_args)

        if data:
            st_in = subprocess.PIPE

        kwargs = dict(
            executable=executable,
            shell=shell,
            close_fds=close_fds,
            stdin=st_in,
            stdout=subprocess.PIPE,
            stderr=subprocess.PIPE,
        )

        if cwd and os.path.isdir(cwd):
            kwargs['cwd'] = cwd

        # store the pwd
        prev_dir = os.getcwd()

        # make sure we're in the right working directory
        if cwd and os.path.isdir(cwd):
            try:
                os.chdir(cwd)
            except (OSError, IOError):
                e = get_exception()
                self.fail_json(rc=e.errno, msg="Could not open %s, %s" % (cwd, str(e)))

        try:

            if self._debug:
                if isinstance(args, list):
                    running = ' '.join(args)
                else:
                    running = args
                self.log('Executing: ' + running)
            cmd = subprocess.Popen(args, **kwargs)

            # the communication logic here is essentially taken from that
            # of the _communicate() function in ssh.py

            stdout = b('')
            stderr = b('')
            rpipes = [cmd.stdout, cmd.stderr]

            if data:
                if not binary_data:
                    data += '\n'
                if isinstance(data, text_type):
                    if PY3:
                        errors = 'surrogateescape'
                    else:
                        errors = 'strict'
                    data = data.encode('utf-8', errors=errors)
                cmd.stdin.write(data)
                cmd.stdin.close()

            while True:
                rfd, wfd, efd = select.select(rpipes, [], rpipes, 1)
                if cmd.stdout in rfd:
                    dat = os.read(cmd.stdout.fileno(), 9000)
                    stdout += dat
                    if dat == b(''):
                        rpipes.remove(cmd.stdout)
                if cmd.stderr in rfd:
                    dat = os.read(cmd.stderr.fileno(), 9000)
                    stderr += dat
                    if dat == b(''):
                        rpipes.remove(cmd.stderr)
                # if we're checking for prompts, do it now
                if prompt_re:
                    if prompt_re.search(stdout) and not data:
                        return (257, stdout, "A prompt was encountered while running a command, but no input data was specified")
                # only break out if no pipes are left to read or
                # the pipes are completely read and
                # the process is terminated
                if (not rpipes or not rfd) and cmd.poll() is not None:
                    break
                # No pipes are left to read but process is not yet terminated
                # Only then it is safe to wait for the process to be finished
                # NOTE: Actually cmd.poll() is always None here if rpipes is empty
                elif not rpipes and cmd.poll() == None:
                    cmd.wait()
                    # The process is terminated. Since no pipes to read from are
                    # left, there is no need to call select() again.
                    break

            cmd.stdout.close()
            cmd.stderr.close()

            rc = cmd.returncode
        except (OSError, IOError):
            e = get_exception()
            self.fail_json(rc=e.errno, msg=str(e), cmd=clean_args)
        except:
            self.fail_json(rc=257, msg=traceback.format_exc(), cmd=clean_args)

        # Restore env settings
        for key, val in old_env_vals.items():
            if val is None:
                del os.environ[key]
            else:
                os.environ[key] = val

        if rc != 0 and check_rc:
            msg = heuristic_log_sanitize(stderr.rstrip(), self.no_log_values)
            self.fail_json(cmd=clean_args, rc=rc, stdout=stdout, stderr=stderr, msg=msg)

        # reset the pwd
        os.chdir(prev_dir)

        return (rc, stdout, stderr)

    def append_to_file(self, filename, str):
        filename = os.path.expandvars(os.path.expanduser(filename))
        fh = open(filename, 'a')
        fh.write(str)
        fh.close()

    def pretty_bytes(self,size):
        ranges = (
                (1<<70, 'ZB'),
                (1<<60, 'EB'),
                (1<<50, 'PB'),
                (1<<40, 'TB'),
                (1<<30, 'GB'),
                (1<<20, 'MB'),
                (1<<10, 'KB'),
                (1, 'Bytes')
            )
        for limit, suffix in ranges:
            if size >= limit:
                break
        return '%.2f %s' % (float(size)/ limit, suffix)

    #
    # Backwards compat
    #

    # In 2.0, moved from inside the module to the toplevel
    is_executable = is_executable


def get_module_path():
    return os.path.dirname(os.path.realpath(__file__))