1088 lines
42 KiB
Python
1088 lines
42 KiB
Python
#!/usr/bin/env python3
|
|
# -*- coding: utf-8 -*-
|
|
|
|
# kindlekey.py
|
|
# Copyright © 2008-2022 Apprentice Harper et al.
|
|
|
|
__license__ = 'GPL v3'
|
|
__version__ = '3.1'
|
|
|
|
# Revision history:
|
|
# 1.0 - Kindle info file decryption, extracted from k4mobidedrm, etc.
|
|
# 1.1 - Added Tkinter to match adobekey.py
|
|
# 1.2 - Fixed testing of successful retrieval on Mac
|
|
# 1.3 - Added getkey interface for Windows DeDRM application
|
|
# Simplified some of the Kindle for Mac code.
|
|
# 1.4 - Remove dependency on alfcrypto
|
|
# 1.5 - moved unicode_argv call inside main for Windows DeDRM compatibility
|
|
# 1.6 - Fixed a problem getting the disk serial numbers
|
|
# 1.7 - Work if TkInter is missing
|
|
# 1.8 - Fixes for Kindle for Mac, and non-ascii in Windows user names
|
|
# 1.9 - Fixes for Unicode in Windows user names
|
|
# 2.0 - Added comments and extra fix for non-ascii Windows user names
|
|
# 2.1 - Fixed Kindle for PC encryption changes March 2016
|
|
# 2.2 - Fixes for Macs with bonded ethernet ports
|
|
# Also removed old .kinfo file support (pre-2011)
|
|
# 2.3 - Added more field names thanks to concavegit's KFX code.
|
|
# 2.4 - Fix for complex Mac disk setups, thanks to Tibs
|
|
# 2.5 - Final Fix for Windows user names with non-ascii characters, thanks to oneofusoneofus
|
|
# 2.6 - Start adding support for Kindle 1.25+ .kinf2018 file
|
|
# 2.7 - Finish .kinf2018 support, PC & Mac by Apprentice Sakuya
|
|
# 2.8 - Fix for Mac OS X Big Sur
|
|
# 3.0 - Python 3 for calibre 5.0
|
|
# 3.1 - Only support PyCryptodome; clean up the code
|
|
|
|
|
|
"""
|
|
Retrieve Kindle for PC/Mac user key.
|
|
"""
|
|
|
|
import sys, os, re
|
|
import codecs
|
|
from struct import pack, unpack, unpack_from
|
|
import json
|
|
import getopt
|
|
import traceback
|
|
import hashlib
|
|
|
|
try:
|
|
from Cryptodome.Cipher import AES
|
|
from Cryptodome.Util import Counter
|
|
from Cryptodome.Protocol.KDF import PBKDF2
|
|
except ImportError:
|
|
from Crypto.Cipher import AES
|
|
from Crypto.Util import Counter
|
|
from Crypto.Protocol.KDF import PBKDF2
|
|
|
|
try:
|
|
RegError
|
|
except NameError:
|
|
class RegError(Exception):
|
|
pass
|
|
|
|
# Routines common to Mac and PC
|
|
|
|
# Wrap a stream so that output gets flushed immediately
|
|
# and also make sure that any unicode strings get
|
|
# encoded using "replace" before writing them.
|
|
class SafeUnbuffered:
|
|
def __init__(self, stream):
|
|
self.stream = stream
|
|
self.encoding = stream.encoding
|
|
if self.encoding == None:
|
|
self.encoding = "utf-8"
|
|
def write(self, data):
|
|
if isinstance(data,str) or isinstance(data,unicode):
|
|
# str for Python3, unicode for Python2
|
|
data = data.encode(self.encoding,"replace")
|
|
try:
|
|
buffer = getattr(self.stream, 'buffer', self.stream)
|
|
# self.stream.buffer for Python3, self.stream for Python2
|
|
buffer.write(data)
|
|
buffer.flush()
|
|
except:
|
|
# We can do nothing if a write fails
|
|
raise
|
|
def __getattr__(self, attr):
|
|
return getattr(self.stream, attr)
|
|
|
|
try:
|
|
from calibre.constants import iswindows, isosx
|
|
except:
|
|
iswindows = sys.platform.startswith('win')
|
|
isosx = sys.platform.startswith('darwin')
|
|
|
|
def unicode_argv():
|
|
if iswindows:
|
|
# Uses shell32.GetCommandLineArgvW to get sys.argv as a list of Unicode
|
|
# strings.
|
|
|
|
# Versions 2.x of Python don't support Unicode in sys.argv on
|
|
# Windows, with the underlying Windows API instead replacing multi-byte
|
|
# characters with '?'. So use shell32.GetCommandLineArgvW to get sys.argv
|
|
# as a list of Unicode strings and encode them as utf-8
|
|
|
|
from ctypes import POINTER, byref, cdll, c_int, windll
|
|
from ctypes.wintypes import LPCWSTR, LPWSTR
|
|
|
|
GetCommandLineW = cdll.kernel32.GetCommandLineW
|
|
GetCommandLineW.argtypes = []
|
|
GetCommandLineW.restype = LPCWSTR
|
|
|
|
CommandLineToArgvW = windll.shell32.CommandLineToArgvW
|
|
CommandLineToArgvW.argtypes = [LPCWSTR, POINTER(c_int)]
|
|
CommandLineToArgvW.restype = POINTER(LPWSTR)
|
|
|
|
cmd = GetCommandLineW()
|
|
argc = c_int(0)
|
|
argv = CommandLineToArgvW(cmd, byref(argc))
|
|
if argc.value > 0:
|
|
# Remove Python executable and commands if present
|
|
start = argc.value - len(sys.argv)
|
|
return [argv[i] for i in
|
|
range(start, argc.value)]
|
|
# if we don't have any arguments at all, just pass back script name
|
|
# this should never happen
|
|
return ["kindlekey.py"]
|
|
else:
|
|
argvencoding = sys.stdin.encoding or "utf-8"
|
|
return [arg if (isinstance(arg, str) or isinstance(arg,unicode)) else str(arg, argvencoding) for arg in sys.argv]
|
|
|
|
class DrmException(Exception):
|
|
pass
|
|
|
|
# crypto digestroutines
|
|
|
|
def MD5(message):
|
|
return hashlib.md5(message).digest()
|
|
|
|
def SHA1(message):
|
|
return hashlib.sha1(message).digest()
|
|
|
|
def SHA256(message):
|
|
return hashlib.sha256(message).digest()
|
|
|
|
|
|
# For K4M/PC 1.6.X and later
|
|
def primes(n):
|
|
"""
|
|
Return a list of prime integers smaller than or equal to n
|
|
:param n: int
|
|
:return: list->int
|
|
"""
|
|
if n == 2:
|
|
return [2]
|
|
elif n < 2:
|
|
return []
|
|
primeList = [2]
|
|
|
|
for potentialPrime in range(3, n + 1, 2):
|
|
isItPrime = True
|
|
for prime in primeList:
|
|
if potentialPrime % prime == 0:
|
|
isItPrime = False
|
|
if isItPrime is True:
|
|
primeList.append(potentialPrime)
|
|
|
|
return primeList
|
|
|
|
# Encode the bytes in data with the characters in map
|
|
# data and map should be byte arrays
|
|
def encode(data, map):
|
|
result = b''
|
|
for char in data:
|
|
value = char
|
|
Q = (value ^ 0x80) // len(map)
|
|
R = value % len(map)
|
|
result += bytes([map[Q]])
|
|
result += bytes([map[R]])
|
|
return result
|
|
|
|
# Hash the bytes in data and then encode the digest with the characters in map
|
|
def encodeHash(data,map):
|
|
return encode(MD5(data),map)
|
|
|
|
# Decode the string in data with the characters in map. Returns the decoded bytes
|
|
def decode(data,map):
|
|
result = b''
|
|
for i in range (0,len(data)-1,2):
|
|
high = map.find(data[i])
|
|
low = map.find(data[i+1])
|
|
if (high == -1) or (low == -1) :
|
|
break
|
|
value = (((high * len(map)) ^ 0x80) & 0xFF) + low
|
|
result += pack('B',value)
|
|
return result
|
|
|
|
def UnprotectHeaderData(encryptedData):
|
|
passwdData = b'header_key_data'
|
|
salt = b'HEADER.2011'
|
|
key_iv = PBKDF2(passwdData, salt, dkLen=256, count=128)
|
|
return AES.new(key_iv[0:32], AES.MODE_CBC, key_iv[32:48]).decrypt(encryptedData)
|
|
|
|
# Routines unique to Mac and PC
|
|
if iswindows:
|
|
from ctypes import windll, c_char_p, c_wchar_p, c_uint, POINTER, byref, \
|
|
create_unicode_buffer, create_string_buffer, CFUNCTYPE, addressof, \
|
|
string_at, Structure, c_void_p, cast
|
|
|
|
try:
|
|
import winreg
|
|
except ImportError:
|
|
import _winreg as winreg
|
|
|
|
MAX_PATH = 255
|
|
kernel32 = windll.kernel32
|
|
advapi32 = windll.advapi32
|
|
crypt32 = windll.crypt32
|
|
|
|
# Various character maps used to decrypt kindle info values.
|
|
# Probably supposed to act as obfuscation
|
|
charMap2 = b"AaZzB0bYyCc1XxDdW2wEeVv3FfUuG4g-TtHh5SsIiR6rJjQq7KkPpL8lOoMm9Nn_"
|
|
charMap5 = b"AzB0bYyCeVvaZ3FfUuG4g-TtHh5SsIiR6rJjQq7KkPpL8lOoMm9Nn_c1XxDdW2wE"
|
|
# New maps in K4PC 1.9.0
|
|
testMap1 = b"n5Pr6St7Uv8Wx9YzAb0Cd1Ef2Gh3Jk4M"
|
|
testMap6 = b"9YzAb0Cd1Ef2n5Pr6St7Uvh3Jk4M8WxG"
|
|
testMap8 = b"YvaZ3FfUm9Nn_c1XuG4yCAzB0beVg-TtHh5SsIiR6rJjQdW2wEq7KkPpL8lOoMxD"
|
|
|
|
# interface with Windows OS Routines
|
|
class DataBlob(Structure):
|
|
_fields_ = [('cbData', c_uint),
|
|
('pbData', c_void_p)]
|
|
DataBlob_p = POINTER(DataBlob)
|
|
|
|
|
|
def GetSystemDirectory():
|
|
GetSystemDirectoryW = kernel32.GetSystemDirectoryW
|
|
GetSystemDirectoryW.argtypes = [c_wchar_p, c_uint]
|
|
GetSystemDirectoryW.restype = c_uint
|
|
def GetSystemDirectory():
|
|
buffer = create_unicode_buffer(MAX_PATH + 1)
|
|
GetSystemDirectoryW(buffer, len(buffer))
|
|
return buffer.value
|
|
return GetSystemDirectory
|
|
GetSystemDirectory = GetSystemDirectory()
|
|
|
|
def GetVolumeSerialNumber():
|
|
GetVolumeInformationW = kernel32.GetVolumeInformationW
|
|
GetVolumeInformationW.argtypes = [c_wchar_p, c_wchar_p, c_uint,
|
|
POINTER(c_uint), POINTER(c_uint),
|
|
POINTER(c_uint), c_wchar_p, c_uint]
|
|
GetVolumeInformationW.restype = c_uint
|
|
def GetVolumeSerialNumber(path = GetSystemDirectory().split('\\')[0] + '\\'):
|
|
vsn = c_uint(0)
|
|
GetVolumeInformationW(path, None, 0, byref(vsn), None, None, None, 0)
|
|
return str(vsn.value)
|
|
return GetVolumeSerialNumber
|
|
GetVolumeSerialNumber = GetVolumeSerialNumber()
|
|
|
|
def GetIDString():
|
|
vsn = GetVolumeSerialNumber()
|
|
#print('Using Volume Serial Number for ID: '+vsn)
|
|
return vsn
|
|
|
|
def getLastError():
|
|
GetLastError = kernel32.GetLastError
|
|
GetLastError.argtypes = None
|
|
GetLastError.restype = c_uint
|
|
def getLastError():
|
|
return GetLastError()
|
|
return getLastError
|
|
getLastError = getLastError()
|
|
|
|
def GetUserName():
|
|
GetUserNameW = advapi32.GetUserNameW
|
|
GetUserNameW.argtypes = [c_wchar_p, POINTER(c_uint)]
|
|
GetUserNameW.restype = c_uint
|
|
def GetUserName():
|
|
buffer = create_unicode_buffer(2)
|
|
size = c_uint(len(buffer))
|
|
while not GetUserNameW(buffer, byref(size)):
|
|
errcd = getLastError()
|
|
if errcd == 234:
|
|
# bad wine implementation up through wine 1.3.21
|
|
return "AlternateUserName"
|
|
# double the buffer size
|
|
buffer = create_unicode_buffer(len(buffer) * 2)
|
|
size.value = len(buffer)
|
|
|
|
# replace any non-ASCII values with 0xfffd
|
|
for i in range(0,len(buffer)):
|
|
if buffer[i]>"\u007f":
|
|
#print "swapping char "+str(i)+" ("+buffer[i]+")"
|
|
buffer[i] = "\ufffd"
|
|
# return utf-8 encoding of modified username
|
|
#print "modified username:"+buffer.value
|
|
return buffer.value.encode('utf-8')
|
|
return GetUserName
|
|
GetUserName = GetUserName()
|
|
|
|
def CryptUnprotectData():
|
|
_CryptUnprotectData = crypt32.CryptUnprotectData
|
|
_CryptUnprotectData.argtypes = [DataBlob_p, c_wchar_p, DataBlob_p,
|
|
c_void_p, c_void_p, c_uint, DataBlob_p]
|
|
_CryptUnprotectData.restype = c_uint
|
|
def CryptUnprotectData(indata, entropy, flags):
|
|
indatab = create_string_buffer(indata)
|
|
indata = DataBlob(len(indata), cast(indatab, c_void_p))
|
|
entropyb = create_string_buffer(entropy)
|
|
entropy = DataBlob(len(entropy), cast(entropyb, c_void_p))
|
|
outdata = DataBlob()
|
|
if not _CryptUnprotectData(byref(indata), None, byref(entropy),
|
|
None, None, flags, byref(outdata)):
|
|
# raise DrmException("Failed to Unprotect Data")
|
|
return b'failed'
|
|
return string_at(outdata.pbData, outdata.cbData)
|
|
return CryptUnprotectData
|
|
CryptUnprotectData = CryptUnprotectData()
|
|
|
|
# Returns Environmental Variables that contain unicode
|
|
# name must be unicode string, not byte string.
|
|
def getEnvironmentVariable(name):
|
|
import ctypes
|
|
n = ctypes.windll.kernel32.GetEnvironmentVariableW(name, None, 0)
|
|
if n == 0:
|
|
return None
|
|
buf = ctypes.create_unicode_buffer("\0"*n)
|
|
ctypes.windll.kernel32.GetEnvironmentVariableW(name, buf, n)
|
|
return buf.value
|
|
|
|
# Locate all of the kindle-info style files and return as list
|
|
def getKindleInfoFiles():
|
|
kInfoFiles = []
|
|
# some 64 bit machines do not have the proper registry key for some reason
|
|
# or the python interface to the 32 vs 64 bit registry is broken
|
|
path = ""
|
|
if 'LOCALAPPDATA' in os.environ.keys():
|
|
# Python 2.x does not return unicode env. Use Python 3.x
|
|
path = winreg.ExpandEnvironmentStrings("%LOCALAPPDATA%")
|
|
# this is just another alternative.
|
|
# path = getEnvironmentVariable('LOCALAPPDATA')
|
|
if not os.path.isdir(path):
|
|
path = ""
|
|
else:
|
|
# User Shell Folders show take precedent over Shell Folders if present
|
|
try:
|
|
# this will still break
|
|
regkey = winreg.OpenKey(winreg.HKEY_CURRENT_USER, "Software\\Microsoft\\Windows\\CurrentVersion\\Explorer\\User Shell Folders\\")
|
|
path = winreg.QueryValueEx(regkey, 'Local AppData')[0]
|
|
if not os.path.isdir(path):
|
|
path = ""
|
|
try:
|
|
regkey = winreg.OpenKey(winreg.HKEY_CURRENT_USER, "Software\\Microsoft\\Windows\\CurrentVersion\\Explorer\\Shell Folders\\")
|
|
path = winreg.QueryValueEx(regkey, 'Local AppData')[0]
|
|
if not os.path.isdir(path):
|
|
path = ""
|
|
except RegError:
|
|
pass
|
|
except RegError:
|
|
pass
|
|
|
|
found = False
|
|
if path == "":
|
|
print ('Could not find the folder in which to look for kinfoFiles.')
|
|
else:
|
|
# Probably not the best. To Fix (shouldn't ignore in encoding) or use utf-8
|
|
print("searching for kinfoFiles in " + path)
|
|
|
|
# look for (K4PC 1.25.1 and later) .kinf2018 file
|
|
kinfopath = path +'\\Amazon\\Kindle\\storage\\.kinf2018'
|
|
if os.path.isfile(kinfopath):
|
|
found = True
|
|
print('Found K4PC 1.25+ kinf2018 file: ' + kinfopath)
|
|
kInfoFiles.append(kinfopath)
|
|
|
|
# look for (K4PC 1.9.0 and later) .kinf2011 file
|
|
kinfopath = path +'\\Amazon\\Kindle\\storage\\.kinf2011'
|
|
if os.path.isfile(kinfopath):
|
|
found = True
|
|
print('Found K4PC 1.9+ kinf2011 file: ' + kinfopath)
|
|
kInfoFiles.append(kinfopath)
|
|
|
|
# look for (K4PC 1.6.0 and later) rainier.2.1.1.kinf file
|
|
kinfopath = path +'\\Amazon\\Kindle\\storage\\rainier.2.1.1.kinf'
|
|
if os.path.isfile(kinfopath):
|
|
found = True
|
|
print('Found K4PC 1.6-1.8 kinf file: ' + kinfopath)
|
|
kInfoFiles.append(kinfopath)
|
|
|
|
# look for (K4PC 1.5.0 and later) rainier.2.1.1.kinf file
|
|
kinfopath = path +'\\Amazon\\Kindle For PC\\storage\\rainier.2.1.1.kinf'
|
|
if os.path.isfile(kinfopath):
|
|
found = True
|
|
print('Found K4PC 1.5 kinf file: ' + kinfopath)
|
|
kInfoFiles.append(kinfopath)
|
|
|
|
# look for original (earlier than K4PC 1.5.0) kindle-info files
|
|
kinfopath = path +'\\Amazon\\Kindle For PC\\{AMAwzsaPaaZAzmZzZQzgZCAkZ3AjA_AY}\\kindle.info'
|
|
if os.path.isfile(kinfopath):
|
|
found = True
|
|
print('Found K4PC kindle.info file: ' + kinfopath)
|
|
kInfoFiles.append(kinfopath)
|
|
|
|
if not found:
|
|
print('No K4PC kindle.info/kinf/kinf2011 files have been found.')
|
|
return kInfoFiles
|
|
|
|
|
|
# determine type of kindle info provided and return a
|
|
# database of keynames and values
|
|
def getDBfromFile(kInfoFile):
|
|
names = [\
|
|
b'kindle.account.tokens',\
|
|
b'kindle.cookie.item',\
|
|
b'eulaVersionAccepted',\
|
|
b'login_date',\
|
|
b'kindle.token.item',\
|
|
b'login',\
|
|
b'kindle.key.item',\
|
|
b'kindle.name.info',\
|
|
b'kindle.device.info',\
|
|
b'MazamaRandomNumber',\
|
|
b'max_date',\
|
|
b'SIGVERIF',\
|
|
b'build_version',\
|
|
b'SerialNumber',\
|
|
b'UsernameHash',\
|
|
b'kindle.directedid.info',\
|
|
b'DSN',\
|
|
b'kindle.accounttype.info',\
|
|
b'krx.flashcardsplugin.data.encryption_key',\
|
|
b'krx.notebookexportplugin.data.encryption_key',\
|
|
b'proxy.http.password',\
|
|
b'proxy.http.username'
|
|
]
|
|
namehashmap = {encodeHash(n,testMap8):n for n in names}
|
|
# print(namehashmap)
|
|
DB = {}
|
|
with open(kInfoFile, 'rb') as infoReader:
|
|
data = infoReader.read()
|
|
# assume .kinf2011 or .kinf2018 style .kinf file
|
|
# the .kinf file uses "/" to separate it into records
|
|
# so remove the trailing "/" to make it easy to use split
|
|
data = data[:-1]
|
|
items = data.split(b'/')
|
|
|
|
# starts with an encoded and encrypted header blob
|
|
headerblob = items.pop(0)
|
|
encryptedValue = decode(headerblob, testMap1)
|
|
cleartext = UnprotectHeaderData(encryptedValue)
|
|
#print "header cleartext:",cleartext
|
|
# now extract the pieces that form the added entropy
|
|
pattern = re.compile(br'''\[Version:(\d+)\]\[Build:(\d+)\]\[Cksum:([^\]]+)\]\[Guid:([\{\}a-z0-9\-]+)\]''', re.IGNORECASE)
|
|
for m in re.finditer(pattern, cleartext):
|
|
version = int(m.group(1))
|
|
build = m.group(2)
|
|
guid = m.group(4)
|
|
|
|
if version == 5: # .kinf2011
|
|
added_entropy = build + guid
|
|
elif version == 6: # .kinf2018
|
|
salt = str(0x6d8 * int(build)).encode('utf-8') + guid
|
|
sp = GetUserName() + b'+@#$%+' + GetIDString().encode('utf-8')
|
|
passwd = encode(SHA256(sp), charMap5)
|
|
key = PBKDF2(passwd, salt, count=10000, dkLen=0x400)[:32] # this is very slow
|
|
|
|
# loop through the item records until all are processed
|
|
while len(items) > 0:
|
|
|
|
# get the first item record
|
|
item = items.pop(0)
|
|
|
|
# the first 32 chars of the first record of a group
|
|
# is the MD5 hash of the key name encoded by charMap5
|
|
keyhash = item[0:32]
|
|
|
|
# the remainder of the first record when decoded with charMap5
|
|
# has the ':' split char followed by the string representation
|
|
# of the number of records that follow
|
|
# and make up the contents
|
|
srcnt = decode(item[34:],charMap5)
|
|
rcnt = int(srcnt)
|
|
|
|
# read and store in rcnt records of data
|
|
# that make up the contents value
|
|
edlst = []
|
|
for i in range(rcnt):
|
|
item = items.pop(0)
|
|
edlst.append(item)
|
|
|
|
# key names now use the new testMap8 encoding
|
|
if keyhash in namehashmap:
|
|
keyname=namehashmap[keyhash]
|
|
#print "keyname found from hash:",keyname
|
|
else:
|
|
keyname = keyhash
|
|
#print "keyname not found, hash is:",keyname
|
|
|
|
# the testMap8 encoded contents data has had a length
|
|
# of chars (always odd) cut off of the front and moved
|
|
# to the end to prevent decoding using testMap8 from
|
|
# working properly, and thereby preventing the ensuing
|
|
# CryptUnprotectData call from succeeding.
|
|
|
|
# The offset into the testMap8 encoded contents seems to be:
|
|
# len(contents)-largest prime number <= int(len(content)/3)
|
|
# (in other words split "about" 2/3rds of the way through)
|
|
|
|
# move first offsets chars to end to align for decode by testMap8
|
|
# by moving noffset chars from the start of the
|
|
# string to the end of the string
|
|
encdata = b"".join(edlst)
|
|
#print "encrypted data:",encdata
|
|
contlen = len(encdata)
|
|
noffset = contlen - primes(int(contlen/3))[-1]
|
|
pfx = encdata[0:noffset]
|
|
encdata = encdata[noffset:]
|
|
encdata = encdata + pfx
|
|
#print "rearranged data:",encdata
|
|
|
|
if version == 5:
|
|
# decode using new testMap8 to get the original CryptProtect Data
|
|
encryptedValue = decode(encdata,testMap8)
|
|
#print "decoded data:",encryptedValue.encode('hex')
|
|
entropy = SHA1(keyhash) + added_entropy
|
|
cleartext = CryptUnprotectData(encryptedValue, entropy, 1)
|
|
elif version == 6:
|
|
# decode using new testMap8 to get IV + ciphertext
|
|
iv_ciphertext = decode(encdata, testMap8)
|
|
# pad IV so that we can substitute AES-CTR for GCM
|
|
iv = iv_ciphertext[:12] + b'\x00\x00\x00\x02'
|
|
ciphertext = iv_ciphertext[12:]
|
|
# convert IV to int for use with pycrypto
|
|
iv_ints = unpack('>QQ', iv)
|
|
iv = iv_ints[0] << 64 | iv_ints[1]
|
|
# set up AES-CTR
|
|
ctr = Counter.new(128, initial_value=iv)
|
|
cipher = AES.new(key, AES.MODE_CTR, counter=ctr)
|
|
# decrypt and decode
|
|
cleartext = decode(cipher.decrypt(ciphertext), charMap5)
|
|
|
|
if len(cleartext)>0:
|
|
#print "cleartext data:",cleartext,":end data"
|
|
DB[keyname] = cleartext
|
|
#print keyname, cleartext
|
|
|
|
if len(DB)>6:
|
|
# store values used in decryption
|
|
DB[b'IDString'] = GetIDString().encode('utf-8')
|
|
DB[b'UserName'] = GetUserName()
|
|
print("Decrypted key file using IDString '{0:s}' and UserName '{1:s}'".format(GetIDString(), GetUserName().decode('utf-8')))
|
|
else:
|
|
print("Couldn't decrypt file.")
|
|
DB = {}
|
|
return DB
|
|
elif isosx:
|
|
import subprocess
|
|
|
|
# Various character maps used to decrypt books. Probably supposed to act as obfuscation
|
|
charMap1 = b'n5Pr6St7Uv8Wx9YzAb0Cd1Ef2Gh3Jk4M'
|
|
charMap2 = b'ZB0bYyc1xDdW2wEV3Ff7KkPpL8UuGA4gz-Tme9Nn_tHh5SvXCsIiR6rJjQaqlOoM'
|
|
|
|
# For kinf approach of K4Mac 1.6.X or later
|
|
# On K4PC charMap5 = 'AzB0bYyCeVvaZ3FfUuG4g-TtHh5SsIiR6rJjQq7KkPpL8lOoMm9Nn_c1XxDdW2wE'
|
|
# For Mac they seem to re-use charMap2 here
|
|
charMap5 = charMap2
|
|
|
|
# new in K4M 1.9.X
|
|
testMap8 = b'YvaZ3FfUm9Nn_c1XuG4yCAzB0beVg-TtHh5SsIiR6rJjQdW2wEq7KkPpL8lOoMxD'
|
|
|
|
# uses a sub process to get the Hard Drive Serial Number using ioreg
|
|
# returns serial numbers of all internal hard drive drives
|
|
def GetVolumesSerialNumbers():
|
|
sernums = []
|
|
sernum = os.getenv('MYSERIALNUMBER')
|
|
if sernum != None:
|
|
sernums.append(sernum.strip())
|
|
cmdline = '/usr/sbin/ioreg -w 0 -r -c AppleAHCIDiskDriver'
|
|
cmdline = cmdline.encode(sys.getfilesystemencoding())
|
|
p = subprocess.Popen(cmdline, shell=True, stdin=None, stdout=subprocess.PIPE, stderr=subprocess.PIPE, close_fds=False)
|
|
out1, out2 = p.communicate()
|
|
#print out1
|
|
reslst = out1.split(b'\n')
|
|
cnt = len(reslst)
|
|
for j in range(cnt):
|
|
resline = reslst[j]
|
|
pp = resline.find(b'\"Serial Number\" = \"')
|
|
if pp >= 0:
|
|
sernum = resline[pp+19:-1]
|
|
sernums.append(sernum.strip())
|
|
return sernums
|
|
|
|
def GetDiskPartitionNames():
|
|
names = []
|
|
cmdline = '/sbin/mount'
|
|
cmdline = cmdline.encode(sys.getfilesystemencoding())
|
|
p = subprocess.Popen(cmdline, shell=True, stdin=None, stdout=subprocess.PIPE, stderr=subprocess.PIPE, close_fds=False)
|
|
out1, out2 = p.communicate()
|
|
reslst = out1.split(b'\n')
|
|
cnt = len(reslst)
|
|
for j in range(cnt):
|
|
resline = reslst[j]
|
|
if resline.startswith(b'/dev'):
|
|
(devpart, mpath) = resline.split(b' on ')[:2]
|
|
dpart = devpart[5:]
|
|
names.append(dpart)
|
|
return names
|
|
|
|
# uses a sub process to get the UUID of all disk partitions
|
|
def GetDiskPartitionUUIDs():
|
|
uuids = []
|
|
uuidnum = os.getenv('MYUUIDNUMBER')
|
|
if uuidnum != None:
|
|
uuids.append(uuidnum.strip())
|
|
cmdline = '/usr/sbin/ioreg -l -S -w 0 -r -c AppleAHCIDiskDriver'
|
|
cmdline = cmdline.encode(sys.getfilesystemencoding())
|
|
p = subprocess.Popen(cmdline, shell=True, stdin=None, stdout=subprocess.PIPE, stderr=subprocess.PIPE, close_fds=False)
|
|
out1, out2 = p.communicate()
|
|
#print out1
|
|
reslst = out1.split(b'\n')
|
|
cnt = len(reslst)
|
|
for j in range(cnt):
|
|
resline = reslst[j]
|
|
pp = resline.find(b'\"UUID\" = \"')
|
|
if pp >= 0:
|
|
uuidnum = resline[pp+10:-1]
|
|
uuidnum = uuidnum.strip()
|
|
uuids.append(uuidnum)
|
|
return uuids
|
|
|
|
def GetMACAddressesMunged():
|
|
macnums = []
|
|
macnum = os.getenv('MYMACNUM')
|
|
if macnum != None:
|
|
macnums.append(macnum)
|
|
cmdline = 'networksetup -listallhardwareports' # en0'
|
|
cmdline = cmdline.encode(sys.getfilesystemencoding())
|
|
p = subprocess.Popen(cmdline, shell=True, stdin=None, stdout=subprocess.PIPE, stderr=subprocess.PIPE, close_fds=False)
|
|
out1, out2 = p.communicate()
|
|
reslst = out1.split(b'\n')
|
|
cnt = len(reslst)
|
|
for j in range(cnt):
|
|
resline = reslst[j]
|
|
pp = resline.find(b'Ethernet Address: ')
|
|
if pp >= 0:
|
|
#print resline
|
|
macnum = resline[pp+18:]
|
|
macnum = macnum.strip()
|
|
maclst = macnum.split(b':')
|
|
n = len(maclst)
|
|
if n != 6:
|
|
continue
|
|
#print 'original mac', macnum
|
|
# now munge it up the way Kindle app does
|
|
# by xoring it with 0xa5 and swapping elements 3 and 4
|
|
for i in range(6):
|
|
maclst[i] = int(b'0x' + maclst[i], 0)
|
|
mlst = [0x00, 0x00, 0x00, 0x00, 0x00, 0x00]
|
|
mlst[5] = maclst[5] ^ 0xa5
|
|
mlst[4] = maclst[3] ^ 0xa5
|
|
mlst[3] = maclst[4] ^ 0xa5
|
|
mlst[2] = maclst[2] ^ 0xa5
|
|
mlst[1] = maclst[1] ^ 0xa5
|
|
mlst[0] = maclst[0] ^ 0xa5
|
|
macnum = b'%0.2x%0.2x%0.2x%0.2x%0.2x%0.2x' % (mlst[0], mlst[1], mlst[2], mlst[3], mlst[4], mlst[5])
|
|
#print 'munged mac', macnum
|
|
macnums.append(macnum)
|
|
return macnums
|
|
|
|
|
|
# uses unix env to get username instead of using sysctlbyname
|
|
def GetUserName():
|
|
username = os.getenv('USER')
|
|
#print "Username:",username
|
|
return username.encode('utf-8')
|
|
|
|
def GetIDStrings():
|
|
# Return all possible ID Strings
|
|
strings = []
|
|
strings.extend(GetMACAddressesMunged())
|
|
strings.extend(GetVolumesSerialNumbers())
|
|
strings.extend(GetDiskPartitionNames())
|
|
strings.extend(GetDiskPartitionUUIDs())
|
|
strings.append(b'9999999999')
|
|
#print "ID Strings:\n",strings
|
|
return strings
|
|
|
|
# implements an Pseudo Mac Version of Windows built-in Crypto routine
|
|
class CryptUnprotectData(object):
|
|
def __init__(self, entropy, IDString):
|
|
sp = GetUserName() + b'+@#$%+' + IDString
|
|
passwdData = encode(SHA256(sp),charMap2)
|
|
salt = entropy
|
|
key_iv = PBKDF2(passwdData, salt, count=0x800, dkLen=0x400)
|
|
self.key = key_iv[0:32]
|
|
self.iv = key_iv[32:48]
|
|
self.crp.set_decrypt_key(self.key, self.iv)
|
|
|
|
def decrypt(self, encryptedData):
|
|
cleartext = self.crp.decrypt(encryptedData)
|
|
cleartext = decode(cleartext, charMap2)
|
|
return cleartext
|
|
|
|
|
|
# Locate the .kindle-info files
|
|
def getKindleInfoFiles():
|
|
# file searches can take a long time on some systems, so just look in known specific places.
|
|
kInfoFiles=[]
|
|
found = False
|
|
home = os.getenv('HOME')
|
|
# check for .kinf2018 file in new location (App Store Kindle for Mac)
|
|
testpath = home + '/Library/Containers/com.amazon.Kindle/Data/Library/Application Support/Kindle/storage/.kinf2018'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kinf2018 file: ' + testpath)
|
|
found = True
|
|
# check for .kinf2018 files
|
|
testpath = home + '/Library/Application Support/Kindle/storage/.kinf2018'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kinf2018 file: ' + testpath)
|
|
found = True
|
|
# check for .kinf2011 file in new location (App Store Kindle for Mac)
|
|
testpath = home + '/Library/Containers/com.amazon.Kindle/Data/Library/Application Support/Kindle/storage/.kinf2011'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kinf2011 file: ' + testpath)
|
|
found = True
|
|
# check for .kinf2011 files from 1.10
|
|
testpath = home + '/Library/Application Support/Kindle/storage/.kinf2011'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kinf2011 file: ' + testpath)
|
|
found = True
|
|
# check for .rainier-2.1.1-kinf files from 1.6
|
|
testpath = home + '/Library/Application Support/Kindle/storage/.rainier-2.1.1-kinf'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac rainier file: ' + testpath)
|
|
found = True
|
|
# check for .kindle-info files from 1.4
|
|
testpath = home + '/Library/Application Support/Kindle/storage/.kindle-info'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kindle-info file: ' + testpath)
|
|
found = True
|
|
# check for .kindle-info file from 1.2.2
|
|
testpath = home + '/Library/Application Support/Amazon/Kindle/storage/.kindle-info'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kindle-info file: ' + testpath)
|
|
found = True
|
|
# check for .kindle-info file from 1.0 beta 1 (27214)
|
|
testpath = home + '/Library/Application Support/Amazon/Kindle for Mac/storage/.kindle-info'
|
|
if os.path.isfile(testpath):
|
|
kInfoFiles.append(testpath)
|
|
print('Found k4Mac kindle-info file: ' + testpath)
|
|
found = True
|
|
if not found:
|
|
print('No k4Mac kindle-info/rainier/kinf2011 files have been found.')
|
|
return kInfoFiles
|
|
|
|
# determine type of kindle info provided and return a
|
|
# database of keynames and values
|
|
def getDBfromFile(kInfoFile):
|
|
names = [\
|
|
b'kindle.account.tokens',\
|
|
b'kindle.cookie.item',\
|
|
b'eulaVersionAccepted',\
|
|
b'login_date',\
|
|
b'kindle.token.item',\
|
|
b'login',\
|
|
b'kindle.key.item',\
|
|
b'kindle.name.info',\
|
|
b'kindle.device.info',\
|
|
b'MazamaRandomNumber',\
|
|
b'max_date',\
|
|
b'SIGVERIF',\
|
|
b'build_version',\
|
|
b'SerialNumber',\
|
|
b'UsernameHash',\
|
|
b'kindle.directedid.info',\
|
|
b'DSN',\
|
|
b'kindle.accounttype.info',\
|
|
b'krx.flashcardsplugin.data.encryption_key',\
|
|
b'krx.notebookexportplugin.data.encryption_key',\
|
|
b'proxy.http.password',\
|
|
b'proxy.http.username'
|
|
]
|
|
with open(kInfoFile, 'rb') as infoReader:
|
|
filedata = infoReader.read()
|
|
|
|
data = filedata[:-1]
|
|
items = data.split(b'/')
|
|
IDStrings = GetIDStrings()
|
|
print ("trying username ", GetUserName(), " on file ", kInfoFile)
|
|
for IDString in IDStrings:
|
|
print ("trying IDString:",IDString)
|
|
try:
|
|
DB = {}
|
|
items = data.split(b'/')
|
|
|
|
# the headerblob is the encrypted information needed to build the entropy string
|
|
headerblob = items.pop(0)
|
|
#print ("headerblob: ",headerblob)
|
|
encryptedValue = decode(headerblob, charMap1)
|
|
#print ("encryptedvalue: ",encryptedValue)
|
|
cleartext = UnprotectHeaderData(encryptedValue)
|
|
#print ("cleartext: ",cleartext)
|
|
|
|
# now extract the pieces in the same way
|
|
pattern = re.compile(br'''\[Version:(\d+)\]\[Build:(\d+)\]\[Cksum:([^\]]+)\]\[Guid:([\{\}a-z0-9\-]+)\]''', re.IGNORECASE)
|
|
for m in re.finditer(pattern, cleartext):
|
|
version = int(m.group(1))
|
|
build = m.group(2)
|
|
guid = m.group(4)
|
|
|
|
#print ("version",version)
|
|
#print ("build",build)
|
|
#print ("guid",guid,"\n")
|
|
|
|
if version == 5: # .kinf2011: identical to K4PC, except the build number gets multiplied
|
|
entropy = str(0x2df * int(build)).encode('utf-8') + guid
|
|
cud = CryptUnprotectData(entropy,IDString)
|
|
#print ("entropy",entropy)
|
|
#print ("cud",cud)
|
|
|
|
elif version == 6: # .kinf2018: identical to K4PC
|
|
salt = str(0x6d8 * int(build)).encode('utf-8') + guid
|
|
sp = GetUserName() + b'+@#$%+' + IDString
|
|
passwd = encode(SHA256(sp), charMap5)
|
|
key = PBKDF2(passwd, salt, count=10000, dkLen=0x400)[:32]
|
|
|
|
#print ("salt",salt)
|
|
#print ("sp",sp)
|
|
#print ("passwd",passwd)
|
|
#print ("key",key)
|
|
|
|
# loop through the item records until all are processed
|
|
while len(items) > 0:
|
|
|
|
# get the first item record
|
|
item = items.pop(0)
|
|
|
|
# the first 32 chars of the first record of a group
|
|
# is the MD5 hash of the key name encoded by charMap5
|
|
keyhash = item[0:32]
|
|
keyname = b'unknown'
|
|
|
|
# unlike K4PC the keyhash is not used in generating entropy
|
|
# entropy = SHA1(keyhash) + added_entropy
|
|
# entropy = added_entropy
|
|
|
|
# the remainder of the first record when decoded with charMap5
|
|
# has the ':' split char followed by the string representation
|
|
# of the number of records that follow
|
|
# and make up the contents
|
|
srcnt = decode(item[34:],charMap5)
|
|
rcnt = int(srcnt)
|
|
|
|
# read and store in rcnt records of data
|
|
# that make up the contents value
|
|
edlst = []
|
|
for i in range(rcnt):
|
|
item = items.pop(0)
|
|
edlst.append(item)
|
|
|
|
keyname = b'unknown'
|
|
for name in names:
|
|
if encodeHash(name,testMap8) == keyhash:
|
|
keyname = name
|
|
break
|
|
if keyname == b'unknown':
|
|
keyname = keyhash
|
|
|
|
# the testMap8 encoded contents data has had a length
|
|
# of chars (always odd) cut off of the front and moved
|
|
# to the end to prevent decoding using testMap8 from
|
|
# working properly, and thereby preventing the ensuing
|
|
# CryptUnprotectData call from succeeding.
|
|
|
|
# The offset into the testMap8 encoded contents seems to be:
|
|
# len(contents) - largest prime number less than or equal to int(len(content)/3)
|
|
# (in other words split 'about' 2/3rds of the way through)
|
|
|
|
# move first offsets chars to end to align for decode by testMap8
|
|
encdata = b''.join(edlst)
|
|
contlen = len(encdata)
|
|
|
|
# now properly split and recombine
|
|
# by moving noffset chars from the start of the
|
|
# string to the end of the string
|
|
noffset = contlen - primes(int(contlen/3))[-1]
|
|
pfx = encdata[0:noffset]
|
|
encdata = encdata[noffset:]
|
|
encdata = encdata + pfx
|
|
|
|
if version == 5:
|
|
# decode using testMap8 to get the CryptProtect Data
|
|
encryptedValue = decode(encdata,testMap8)
|
|
cleartext = cud.decrypt(encryptedValue)
|
|
|
|
elif version == 6:
|
|
# decode using new testMap8 to get IV + ciphertext
|
|
iv_ciphertext = decode(encdata, testMap8)
|
|
# pad IV so that we can substitute AES-CTR for GCM
|
|
iv = iv_ciphertext[:12] + b'\x00\x00\x00\x02'
|
|
ciphertext = iv_ciphertext[12:]
|
|
# convert IV to int for use with pycrypto
|
|
iv_ints = unpack('>QQ', iv)
|
|
iv = iv_ints[0] << 64 | iv_ints[1]
|
|
# set up AES-CTR
|
|
ctr = Counter.new(128, initial_value=iv)
|
|
cipher = AES.new(key, AES.MODE_CTR, counter=ctr)
|
|
# decrypt and decode
|
|
cleartext = decode(cipher.decrypt(ciphertext), charMap5)
|
|
|
|
# print keyname
|
|
# print cleartext
|
|
if len(cleartext) > 0:
|
|
DB[keyname] = cleartext
|
|
|
|
if len(DB)>6:
|
|
break
|
|
|
|
except Exception:
|
|
print (traceback.format_exc())
|
|
pass
|
|
if len(DB)>6:
|
|
# store values used in decryption
|
|
print("Decrypted key file using IDString '{0:s}' and UserName '{1:s}'".format(IDString.decode('utf-8'), GetUserName().decode('utf-8')))
|
|
DB[b'IDString'] = IDString
|
|
DB[b'UserName'] = GetUserName()
|
|
else:
|
|
print("Couldn't decrypt file.")
|
|
DB = {}
|
|
return DB
|
|
else:
|
|
def getDBfromFile(kInfoFile):
|
|
raise DrmException("This script only runs under Windows or Mac OS X.")
|
|
return {}
|
|
|
|
def kindlekeys(files = []):
|
|
keys = []
|
|
if files == []:
|
|
files = getKindleInfoFiles()
|
|
for file in files:
|
|
key = getDBfromFile(file)
|
|
if key:
|
|
# convert all values to hex, just in case.
|
|
n_key = {}
|
|
for k,v in key.items():
|
|
n_key[k.decode()]=codecs.encode(v, 'hex_codec').decode()
|
|
# key = {k.decode():v.decode() for k,v in key.items()}
|
|
keys.append(n_key)
|
|
return keys
|
|
|
|
# interface for Python DeDRM
|
|
# returns single key or multiple keys, depending on path or file passed in
|
|
def getkey(outpath, files=[]):
|
|
keys = kindlekeys(files)
|
|
if len(keys) > 0:
|
|
if not os.path.isdir(outpath):
|
|
outfile = outpath
|
|
with open(outfile, 'w') as keyfileout:
|
|
keyfileout.write(json.dumps(keys[0]))
|
|
print("Saved a key to {0}".format(outfile))
|
|
else:
|
|
keycount = 0
|
|
for key in keys:
|
|
while True:
|
|
keycount += 1
|
|
outfile = os.path.join(outpath,"kindlekey{0:d}.k4i".format(keycount))
|
|
if not os.path.exists(outfile):
|
|
break
|
|
with open(outfile, 'w') as keyfileout:
|
|
keyfileout.write(json.dumps(key))
|
|
print("Saved a key to {0}".format(outfile))
|
|
return True
|
|
return False
|
|
|
|
def usage(progname):
|
|
print("Finds, decrypts and saves the default Kindle For Mac/PC encryption keys.")
|
|
print("Keys are saved to the current directory, or a specified output directory.")
|
|
print("If a file name is passed instead of a directory, only the first key is saved, in that file.")
|
|
print("Usage:")
|
|
print(" {0:s} [-h] [-k <kindle.info>] [<outpath>]".format(progname))
|
|
|
|
|
|
def cli_main():
|
|
sys.stdout=SafeUnbuffered(sys.stdout)
|
|
sys.stderr=SafeUnbuffered(sys.stderr)
|
|
argv=unicode_argv()
|
|
progname = os.path.basename(argv[0])
|
|
print("{0} v{1}\nCopyright © 2010-2020 by some_updates, Apprentice Harper et al.".format(progname,__version__))
|
|
|
|
try:
|
|
opts, args = getopt.getopt(argv[1:], "hk:")
|
|
except getopt.GetoptError as err:
|
|
print("Error in options or arguments: {0}".format(err.args[0]))
|
|
usage(progname)
|
|
sys.exit(2)
|
|
|
|
files = []
|
|
for o, a in opts:
|
|
if o == "-h":
|
|
usage(progname)
|
|
sys.exit(0)
|
|
if o == "-k":
|
|
files = [a]
|
|
|
|
if len(args) > 1:
|
|
usage(progname)
|
|
sys.exit(2)
|
|
|
|
if len(args) == 1:
|
|
# save to the specified file or directory
|
|
outpath = args[0]
|
|
if not os.path.isabs(outpath):
|
|
outpath = os.path.abspath(outpath)
|
|
else:
|
|
# save to the same directory as the script
|
|
outpath = os.path.dirname(argv[0])
|
|
|
|
# make sure the outpath is canonical
|
|
outpath = os.path.realpath(os.path.normpath(outpath))
|
|
|
|
if not getkey(outpath, files):
|
|
print("Could not retrieve Kindle for Mac/PC key.")
|
|
return 0
|
|
|
|
|
|
def gui_main():
|
|
try:
|
|
import tkinter
|
|
import tkinter.constants
|
|
import tkinter.messagebox
|
|
import traceback
|
|
except:
|
|
return cli_main()
|
|
|
|
class ExceptionDialog(tkinter.Frame):
|
|
def __init__(self, root, text):
|
|
tkinter.Frame.__init__(self, root, border=5)
|
|
label = tkinter.Label(self, text="Unexpected error:",
|
|
anchor=tkinter.constants.W, justify=tkinter.constants.LEFT)
|
|
label.pack(fill=tkinter.constants.X, expand=0)
|
|
self.text = tkinter.Text(self)
|
|
self.text.pack(fill=tkinter.constants.BOTH, expand=1)
|
|
|
|
self.text.insert(tkinter.constants.END, text)
|
|
|
|
|
|
argv=unicode_argv()
|
|
root = tkinter.Tk()
|
|
root.withdraw()
|
|
progpath, progname = os.path.split(argv[0])
|
|
success = False
|
|
try:
|
|
keys = kindlekeys()
|
|
keycount = 0
|
|
for key in keys:
|
|
while True:
|
|
keycount += 1
|
|
outfile = os.path.join(progpath,"kindlekey{0:d}.k4i".format(keycount))
|
|
if not os.path.exists(outfile):
|
|
break
|
|
|
|
with open(outfile, 'w') as keyfileout:
|
|
keyfileout.write(json.dumps(key))
|
|
success = True
|
|
tkinter.messagebox.showinfo(progname, "Key successfully retrieved to {0}".format(outfile))
|
|
except DrmException as e:
|
|
tkinter.messagebox.showerror(progname, "Error: {0}".format(str(e)))
|
|
except Exception:
|
|
root.wm_state('normal')
|
|
root.title(progname)
|
|
text = traceback.format_exc()
|
|
ExceptionDialog(root, text).pack(fill=tkinter.constants.BOTH, expand=1)
|
|
root.mainloop()
|
|
if not success:
|
|
return 1
|
|
return 0
|
|
|
|
if __name__ == '__main__':
|
|
if len(sys.argv) > 1:
|
|
sys.exit(cli_main())
|
|
sys.exit(gui_main())
|