# -*- coding: utf-8 -*- ############################################################################## # # Copyright (c) 2010-2014 Vifib SARL and Contributors. # All Rights Reserved. # # WARNING: This program as such is intended to be used by professional # programmers who take the whole responsibility of assessing all potential # consequences resulting from its eventual inadequacies and bugs # End users who are looking for a ready-to-use solution with commercial # guarantees and support are strongly adviced to contract a Free Software # Service Company # # This program is Free Software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public License # as published by the Free Software Foundation; either version 2.1 # of the License, or (at your option) any later version. # # 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 Lesser General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. # ############################################################################## import enum import errno import hashlib import json import os import pprint import shutil import socket import sqlite3 import struct import subprocess import warnings import jsonschema import netaddr import requests import six from lxml import etree from six.moves.urllib import parse from six.moves.urllib_parse import urljoin from xml_marshaller.xml_marshaller import Marshaller, Unmarshaller try: from typing import Dict, Optional, IO except ImportError: pass try: PermissionError except NameError: # make pylint happy on python2... PermissionError = Exception _ALLOWED_CLASS_SET = frozenset(( ('slapos.slap.slap', 'Computer'), ('slapos.slap.slap', 'ComputerPartition'), ('slapos.slap.slap', 'SoftwareRelease'), ('slapos.slap.slap', 'SoftwareInstance'), )) class UndefinedSerializationError(ValueError): """Raised when the serialization type is not found""" pass class SafeXMLMarshaller(Marshaller): def m_instance(self, value, kw): cls = value.__class__ if (cls.__module__, cls.__name__) in _ALLOWED_CLASS_SET: return super(SafeXMLMarshaller, self).m_instance(value, kw) raise RuntimeError("Refusing to marshall {}.{}".format( cls.__module__, cls.__name__)) dumps = SafeXMLMarshaller().dumps class SafeXMLUnmrshaller(Unmarshaller, object): def find_class(self, module, name): if (module, name) in _ALLOWED_CLASS_SET: return super(SafeXMLUnmrshaller, self).find_class(module, name) raise RuntimeError("Refusing to unmarshall {}.{}".format(module, name)) loads = SafeXMLUnmrshaller().loads def mkdir_p(path, mode=0o700): """\ Creates a directory and its parents, if needed. NB: If the directory already exists, it does not change its permission. """ try: os.makedirs(path, mode) except OSError as exc: if exc.errno == errno.EEXIST and os.path.isdir(path): pass else: raise def listifdir(path): """ Like listdir, but returns an empty tuple if the path is not a directory. """ try: return os.listdir(path) except OSError as e: if e.errno != errno.ENOENT: raise return () def chownDirectory(path, uid, gid): if os.getuid() != 0: # we are probably inside of a webrunner return # find /opt/slapgrid -not -user 1000 -exec chown slapsoft:slapsoft {} + subprocess.check_call([ '/usr/bin/find', path, '-not', '-user', str(uid), '-exec', '/bin/chown', '%s:%s' % (uid, gid), '{}', '+' ]) def parse_certificate_key_pair(html): """ Extract (certificate, key) pair from an HTML page received by SlapOS Master. """ c_start = html.find("Certificate:") c_end = html.find("</textarea>", c_start) certificate = html[c_start:c_end] k_start = html.find("-----BEGIN PRIVATE KEY-----") k_end = html.find("</textarea>", k_start) key = html[k_start:k_end] return certificate, key def string_to_boolean(string): """ Return True if the value of the "string" parameter can be parsed as True. Return False if the value of the "string" parameter can be parsed as False. Otherwise, Raise. The parser is completely arbitrary, see code for actual implementation. """ try: return ('false', 'true').index(string.lower()) except Exception: raise ValueError('%s is neither True nor False.' % string) def sqlite_connect(dburi, timeout=None, isolation_level=None): connect_kw = {} if timeout is not None: connect_kw['timeout'] = timeout connect_kw['isolation_level'] = isolation_level conn = sqlite3.connect(dburi, **connect_kw) conn.text_factory = str # allow 8-bit strings return conn # The 3 functions below were imported from re6st: # https://lab.nexedi.com/nexedi/re6stnet/blob/master/re6st/utils.py def binFromRawIpv6(ip): ip1, ip2 = struct.unpack('>QQ', ip) return bin(ip1)[2:].rjust(64, '0') + bin(ip2)[2:].rjust(64, '0') def binFromIpv6(ip): """ convert an IPv6 to a 128 characters string containing 0 and 1 e.g.: '2001:db8::'-> '001000000000000100001101101110000000000...000' """ return binFromRawIpv6(socket.inet_pton(socket.AF_INET6, ip)) def ipv6FromBin(ip, suffix=''): """ convert a string containing 0 and 1 to an IPv6 if the string is less than 128 characters: * consider the string is the first bits * optionnaly can replace the last bits of the IP with a suffix (in binary string format) """ suffix_len = 128 - len(ip) if suffix_len > 0: ip += suffix.rjust(suffix_len, '0') elif suffix_len: sys.exit("Prefix exceeds 128 bits") return socket.inet_ntop(socket.AF_INET6, struct.pack('>QQ', int(ip[:64], 2), int(ip[64:], 2))) def getPartitionIpv6Addr(ipv6_range, partition_index): """ from a IPv6 range in the form { 'addr' : addr, 'prefixlen' : CIDR } returns the IPv6 addr addr::(partition_index+2) (address 1 is is used by re6st) """ addr = ipv6_range['addr'] prefixlen = ipv6_range['prefixlen'] prefix = binFromIpv6(addr)[:prefixlen] return dict(addr=ipv6FromBin(prefix + bin(partition_index+2)[2:].zfill(128 - prefixlen)), prefixlen=prefixlen) def getIpv6RangeFactory(k, s): def getIpv6Range(ipv6_range, partition_index): """ from a IPv6 range in the form { 'addr' : addr, 'prefixlen' : CIDR } returns the IPv6 range { 'addr' : addr:(k*(2^14) + partition_index+1) 'prefixlen' : CIDR+16 } """ addr = ipv6_range['addr'] prefixlen = ipv6_range['prefixlen'] prefix = binFromIpv6(addr)[:prefixlen] # we generate a subnetwork for the partition # the subnetwork has 16 bits more than our IPv6 range # make sure we have at least 2 IPs in the subnetwork prefixlen += 16 if prefixlen >= 128: raise ValueError('The IPv6 range has prefixlen {} which is too big for generating IPv6 range for partitions.'.format(prefixlen)) return dict(addr=ipv6FromBin(prefix + bin((k << 14) + partition_index+1)[2:].zfill(16) + s * (128 - prefixlen)), prefixlen=prefixlen) return getIpv6Range getPartitionIpv6Range = getIpv6RangeFactory(1, '0') getTapIpv6Range = getIpv6RangeFactory(2, '1') getTunIpv6Range = getIpv6RangeFactory(3, '0') def getIpv6RangeFirstAddr(addr, prefixlen): addr_1 = "%s1" % ipv6FromBin(binFromIpv6(addr)[:prefixlen]) return ipv6FromBin(binFromIpv6(addr_1)) # correctly format the IPv6 def lenNetmaskIpv6(netmask): """Convert string represented netmask to its integer prefix""" # Since version 0.10.7 of netifaces, the netmask is something like "ffff::/16", # (it used to be "ffff::"). For old versions of netifaces, interpret the netmask # as an address and return its netmask, but for newer versions returns the prefixlen. try: return netaddr.IPAddress(netmask).netmask_bits() except ValueError: return netaddr.IPNetwork(netmask).prefixlen def netmaskFromLenIPv6(netmask_len): """ opposite of lenNetmaskIpv6""" return ipv6FromBin('1' * netmask_len) # Used for Python 2-3 compatibility if str is bytes: bytes2str = str2bytes = lambda s: s def unicode2str(s): return s.encode('utf-8') else: def bytes2str(s): return s.decode() def str2bytes(s): return s.encode() def unicode2str(s): return s def dict2xml(dictionary): instance = etree.Element('instance') for k, v in sorted(six.iteritems(dictionary)): if isinstance(k, bytes): k = k.decode('utf-8') if isinstance(v, bytes): v = v.decode('utf-8') elif not isinstance(v, six.text_type): v = str(v) etree.SubElement(instance, "parameter", attrib={'id': k}).text = v return bytes2str(etree.tostring(instance, pretty_print=True, xml_declaration=True, encoding='utf-8')) def xml2dict(xml): result_dict = {} if xml: tree = etree.fromstring(str2bytes(xml)) for element in tree.iterfind('parameter'): key = element.get('id') value = result_dict.get(key, None) if value is not None: value = value + ' ' + element.text else: value = element.text result_dict[key] = value return result_dict def calculate_dict_hash(d): return hashlib.sha256( str2bytes(str( sorted( d.items() ) ))).hexdigest() def _addIpv6Brackets(url): # if master_url contains an ipv6 without bracket, add it # Note that this is mostly to limit specific issues with # backward compatiblity, not to ensure generic detection. api_scheme, api_netloc, api_path, api_query, api_fragment = parse.urlsplit(url) try: ip = netaddr.IPAddress(api_netloc) port = None except netaddr.AddrFormatError: try: ip = netaddr.IPAddress(':'.join(api_netloc.split(':')[:-1])) port = api_netloc.split(':')[-1] except netaddr.AddrFormatError: ip = port = None if ip and ip.version == 6: api_netloc = '[%s]' % ip if port: api_netloc = '%s:%s' % (api_netloc, port) url = parse.urlunsplit((api_scheme, api_netloc, api_path, api_query, api_fragment)) return url def rmtree(path): """Delete a path recursively. Like shutil.rmtree, but supporting the case that some files or folder might have been marked read only. """ def chmod_retry(func, failed_path, exc_info): """Make sure the directories are executable and writable. """ # Depending on the Python version, the following items differ. if six.PY3: expected_error_type = PermissionError expected_func_set = {os.lstat, os.open} else: expected_error_type = OSError expected_func_set = {os.listdir} e = exc_info[1] if isinstance(e, expected_error_type): if e.errno == errno.ENOENT: # because we are calling again rmtree on listdir errors, this path might # have been already deleted by the recursive call to rmtree. return if e.errno == errno.EACCES: if func in expected_func_set: os.chmod(failed_path, 0o700) # corner case to handle errors in listing directories. # https://bugs.python.org/issue8523 return shutil.rmtree(failed_path, onerror=chmod_retry) # If parent directory is not writable, we still cannot delete the file. # But make sure not to change the parent of the folder we are deleting. if failed_path != path: os.chmod(os.path.dirname(failed_path), 0o700) return func(failed_path) raise e # XXX make pylint happy shutil.rmtree(path, onerror=chmod_retry) def _readAsJson(url): # type: (str) -> Optional[Dict] """Reads and parse the json file located at `url`. `url` can also be the path of a local file. """ if url.startswith('file://'): url = url[len('file://'):] path = url if os.path.exists(url) else None if path: with open(path) as f: try: return json.load(f) except ValueError: return None if url.startswith('http://') or url.startswith('https://'): try: r = requests.get(url) r.raise_for_status() return r.json() except (requests.exceptions.RequestException, ValueError): return None return None class SoftwareReleaseSerialisation(str, enum.Enum): Xml = 'xml' JsonInXml = 'json-in-xml' class SoftwareReleaseSchema(object): def __init__(self, software_url, software_type): # type: (str, Optional[str]) -> None self.software_url = software_url self.software_type = software_type def getSoftwareSchema(self): # type: () -> Optional[Dict] """Returns the schema for this software. """ return _readAsJson(self.software_url + '.json') def getSoftwareTypeSchema(self): # type: () -> Optional[Dict] """Returns schema for this software type. """ software_schema = self.getSoftwareSchema() if software_schema is None: return None software_type = self.software_type from slapos.slap.slap import DEFAULT_SOFTWARE_TYPE if software_type is None: software_type = DEFAULT_SOFTWARE_TYPE # XXX Some software are using "default" for default software type, because # we are transitionning from DEFAULT_SOFTWARE_TYPE ("RootSoftwareInstance") # to "default". if software_type == DEFAULT_SOFTWARE_TYPE \ and software_type not in software_schema['software-type'] \ and 'default' in software_schema['software-type']: warnings.warn( "Software release {} does not have schema for DEFAULT_SOFTWARE_TYPE but has one for 'default'." " Using 'default' instead.".format(self.software_url), UserWarning, ) software_type = 'default' return software_schema['software-type'].get(software_type) def getSerialisation(self): # type: () -> SoftwareReleaseSerialisation """Returns the serialisation method used for parameters. """ software_schema = self.getSoftwareTypeSchema() if software_schema is None or 'serialisation' not in software_schema: software_schema = self.getSoftwareSchema() if software_schema is None: raise UndefinedSerializationError return SoftwareReleaseSerialisation(software_schema['serialisation']) def getInstanceRequestParameterSchemaURL(self): # type: () -> Optional[str] """Returns the URL of the schema defining instance parameters. """ software_type_schema = self.getSoftwareTypeSchema() if software_type_schema is None: return None software_url = self.software_url if os.path.exists(software_url): software_url = 'file://' + software_url return urljoin(software_url, software_type_schema['request']) def getInstanceRequestParameterSchema(self): # type: () -> Optional[Dict] """Returns the schema defining instance parameters. """ instance_parameter_schema_url = self.getInstanceRequestParameterSchemaURL() if instance_parameter_schema_url is None: return None schema = _readAsJson(instance_parameter_schema_url) if schema: # so that jsonschema knows how to resolve references schema.setdefault('$id', instance_parameter_schema_url) return schema def getInstanceConnectionParameterSchemaURL(self): # type: () -> Optional[str] """Returns the URL of the schema defining connection parameters published by the instance. """ software_type_schema = self.getSoftwareTypeSchema() if software_type_schema is None: return None return urljoin(self.software_url, software_type_schema['response']) def getInstanceConnectionParameterSchema(self): # type: () -> Optional[Dict] """Returns the schema defining connection parameters published by the instance. """ instance_parameter_schema_url = self.getInstanceConnectionParameterSchemaURL() if instance_parameter_schema_url is None: return None schema = _readAsJson(instance_parameter_schema_url) if schema: # so that jsonschema knows how to resolve references schema.setdefault('$id', instance_parameter_schema_url) return schema def validateInstanceParameterDict(self, parameter_dict): # type: (Dict) -> None """Validate instance parameters against the software schema. Raise jsonschema.ValidationError if parameters does not validate. """ schema_url = self.getInstanceRequestParameterSchemaURL() if schema_url: instance = parameter_dict if self.getSerialisation() == SoftwareReleaseSerialisation.JsonInXml: try: instance = json.loads(parameter_dict['_']) except KeyError: instance = parameter_dict instance.pop('$schema', None) jsonschema.validate( instance=instance, schema=self.getInstanceRequestParameterSchema(), ) # BBB on python3 we can use pprint.pformat class StrPrettyPrinter(pprint.PrettyPrinter): """A PrettyPrinter which produces consistent output on python 2 and 3 """ def format(self, object, context, maxlevels, level): if six.PY2 and isinstance(object, six.text_type): object = object.encode('utf-8') return pprint.PrettyPrinter.format(self, object, context, maxlevels, level)