"--dump_ascii", help="Dump the ascii output from packets ", action="store_true", default=False) parser.add_argument("-f", "--force", help="Take all default options", action="store_true", default=False) args = parser.parse_args() inputFilePath = args.pcap_file # This stores all the fuzzer data and will eventually write it to the .fuzzer file with comments fuzzerData = FuzzerData() fuzzerData.processorDirectory = args.processor_dir[0] ############# Process input files if not os.path.isfile(inputFilePath): print("Cannot read input %s" % (inputFilePath)) exit() STATE_BETWEEN_MESSAGES = 0 STATE_READING_MESSAGE = 2 STATE_COMBINING_PACKETS = 3 data = [] defaultPort = None with open(inputFilePath, 'r') as inputFile: # This is a little naive, but it works
if args.action != "bin2fuzzer" and args.fuzzerfile: print("Use --fuzzerfile with only the bin2fuzzer option, to populate .fuzzer data") exit(1) # Default file descriptors inFileDesc = sys.stdin outFileDesc = sys.stdout # If we get file paths instead, fix them up # Have to do outfile lower down or we'll blow away an output file we might be fixing up if args.infile: inFileDesc = open(args.infile, "r") if args.action == "list": fuzzerData = FuzzerData() # Allow a non-quiet read to list out messages fuzzerData.readFromFD(inFileDesc, quiet=False) elif args.action in ["fuzzer2bin", "bin2fuzzer"]: if args.messagenum == None: print(("Message number required for action {0}".format(args.action))) exit(1) fuzzerData = FuzzerData() if args.action == "fuzzer2bin": # Pull message out from .fuzzer file, output as binary fuzzerData.readFromFD(inFileDesc, quiet=True) messageCount = len(fuzzerData.messageCollection.messages)
def __init__(self, args): self.args = args # Test number to start from, 0 default self.MIN_RUN_NUMBER = 0 # Test number to go to, -1 is unlimited self.MAX_RUN_NUMBER = -1 # For seed loop, finite range to repeat self.SEED_LOOP = [] # For dumpraw option, dump into log directory by default, else 'dumpraw' self.DUMPDIR = "" #Populate global arguments from parseargs self.fuzzerFilePath = args.prepped_fuzz self.host = args.target_host #Assign Lower/Upper bounds on test cases as needed if args.range: (self.MIN_RUN_NUMBER, self.MAX_RUN_NUMBER) = getRunNumbersFromArgs(args, args.range) elif args.loop: self.SEED_LOOP = validateNumberRange(args.loop, True) #Check for dependency binaries if not os.path.exists(RADAMSA): sys.exit("Could not find radamsa in %s... did you build it?" % RADAMSA) #Logging options self.isReproduce = False self.logAll = False if args.quiet: self.isReproduce = True elif args.logAll: self.logAll = True self.outputDataFolderPath = os.path.join( "%s_%s" % (os.path.splitext(self.fuzzerFilePath)[0], "logs"), datetime.datetime.now().strftime("%Y-%m-%d,%H%M%S")) self.fuzzerFolder = os.path.abspath( os.path.dirname(self.fuzzerFilePath)) ########## Declare variables for scoping, "None"s will be assigned below self.messageProcessor = None ###Here we read in the fuzzer file into a dictionary for easier variable propagation self.optionDict = {"unfuzzedBytes": {}, "message": []} self.fuzzerData = FuzzerData() print "Reading in fuzzer data from %s..." % (self.fuzzerFilePath) self.fuzzerData.readFromFile(self.fuzzerFilePath) #clumsden TODO - make this pretty, maybe add field to fuzzerData #clumsden - hacky way to start http fuzzer at a different seed if self.fuzzerData.port == 80 and self.fuzzerData.proto == "tcp": self.MIN_RUN_NUMBER = 58247530 if self.fuzzerData.proto == "udp": #clumsden, set udp fuzzer at different seed self.MIN_RUN_NUMBER = 13243441 #clumsden end ######## Processor Setup ################ # The processor just acts as a container # # class that will import custom versions # # messageProcessor/exceptionProessor/ # # monitor, if they are found in the # # process_dir specified in the .fuzzer # # file generated by fuzz_prep.py # ########################################## # Assign options to variables, error on anything that's missing/invalid self.processorDirectory = self.fuzzerData.processorDirectory if self.processorDirectory == "default": # Default to fuzzer file folder self.processorDirectory = self.fuzzerFolder else: # Make sure fuzzer file path is prepended self.processorDirectory = os.path.join(self.fuzzerFolder, self.processorDirectory) #Create class director, which import/overrides processors as appropriate self.procDirector = ProcDirector(self.processorDirectory) global global_monitor self.monitor = None ########## Launch child monitor thread ### monitor.task = spawned thread ### monitor.crashEvent = threading.Event() if wantGlobalMonitor == True and global_monitor == None: global_monitor = self.procDirector.startMonitor( self.host, self.fuzzerData.port) print global_monitor time.sleep( 10 ) #clumsden added so pid_watcher has time to connect to monitor #! make it so logging message does not appear if reproducing (i.e. -r x-y cmdline arg is set) self.logger = None if not self.isReproduce: print "Logging to %s" % (self.outputDataFolderPath) self.logger = Logger(self.outputDataFolderPath) if self.args.dumpraw: if not isReproduce: self.DUMPDIR = self.outputDataFolderPath else: self.DUMPDIR = "dumpraw" try: os.mkdir("dumpraw") except: print "Unable to create dumpraw dir" pass self.exceptionProcessor = self.procDirector.exceptionProcessor() self.messageProcessor = self.procDirector.messageProcessor() ########## Begin fuzzing self.i = self.MIN_RUN_NUMBER - 1 if self.fuzzerData.shouldPerformTestRun else self.MIN_RUN_NUMBER self.failureCount = 0 self.loop_len = len(self.SEED_LOOP) # if --loop
elif args.logAll: logAll = True outputDataFolderPath = os.path.join( "%s_%s" % (os.path.splitext(fuzzerFilePath)[0], "logs"), datetime.datetime.now().strftime("%Y-%m-%d,%H%M%S")) fuzzerFolder = os.path.abspath(os.path.dirname(fuzzerFilePath)) ########## Declare variables for scoping, "None"s will be assigned below messageProcessor = None monitor = None ###Here we read in the fuzzer file into a dictionary for easier variable propagation optionDict = {"unfuzzedBytes": {}, "message": []} fuzzerData = FuzzerData() print "Reading in fuzzer data from %s..." % (fuzzerFilePath) fuzzerData.readFromFile(fuzzerFilePath) ######## Processor Setup ################ # The processor just acts as a container # # class that will import custom versions # # messageProcessor/exceptionProessor/ # # monitor, if they are found in the # # process_dir specified in the .fuzzer # # file generated by fuzz_prep.py # ########################################## # Assign options to variables, error on anything that's missing/invalid processorDirectory = fuzzerData.processorDirectory if processorDirectory == "default":
class MutinyFuzzer(): def __init__(self, args): self.args = args # Test number to start from, 0 default self.MIN_RUN_NUMBER = 0 # Test number to go to, -1 is unlimited self.MAX_RUN_NUMBER = -1 # For seed loop, finite range to repeat self.SEED_LOOP = [] # For dumpraw option, dump into log directory by default, else 'dumpraw' self.DUMPDIR = "" #Populate global arguments from parseargs self.fuzzerFilePath = args.prepped_fuzz self.host = args.target_host #Assign Lower/Upper bounds on test cases as needed if args.range: (self.MIN_RUN_NUMBER, self.MAX_RUN_NUMBER) = getRunNumbersFromArgs(args, args.range) elif args.loop: self.SEED_LOOP = validateNumberRange(args.loop, True) #Check for dependency binaries if not os.path.exists(RADAMSA): sys.exit("Could not find radamsa in %s... did you build it?" % RADAMSA) #Logging options self.isReproduce = False self.logAll = False if args.quiet: self.isReproduce = True elif args.logAll: self.logAll = True self.outputDataFolderPath = os.path.join( "%s_%s" % (os.path.splitext(self.fuzzerFilePath)[0], "logs"), datetime.datetime.now().strftime("%Y-%m-%d,%H%M%S")) self.fuzzerFolder = os.path.abspath( os.path.dirname(self.fuzzerFilePath)) ########## Declare variables for scoping, "None"s will be assigned below self.messageProcessor = None ###Here we read in the fuzzer file into a dictionary for easier variable propagation self.optionDict = {"unfuzzedBytes": {}, "message": []} self.fuzzerData = FuzzerData() print "Reading in fuzzer data from %s..." % (self.fuzzerFilePath) self.fuzzerData.readFromFile(self.fuzzerFilePath) #clumsden TODO - make this pretty, maybe add field to fuzzerData #clumsden - hacky way to start http fuzzer at a different seed if self.fuzzerData.port == 80 and self.fuzzerData.proto == "tcp": self.MIN_RUN_NUMBER = 58247530 if self.fuzzerData.proto == "udp": #clumsden, set udp fuzzer at different seed self.MIN_RUN_NUMBER = 13243441 #clumsden end ######## Processor Setup ################ # The processor just acts as a container # # class that will import custom versions # # messageProcessor/exceptionProessor/ # # monitor, if they are found in the # # process_dir specified in the .fuzzer # # file generated by fuzz_prep.py # ########################################## # Assign options to variables, error on anything that's missing/invalid self.processorDirectory = self.fuzzerData.processorDirectory if self.processorDirectory == "default": # Default to fuzzer file folder self.processorDirectory = self.fuzzerFolder else: # Make sure fuzzer file path is prepended self.processorDirectory = os.path.join(self.fuzzerFolder, self.processorDirectory) #Create class director, which import/overrides processors as appropriate self.procDirector = ProcDirector(self.processorDirectory) global global_monitor self.monitor = None ########## Launch child monitor thread ### monitor.task = spawned thread ### monitor.crashEvent = threading.Event() if wantGlobalMonitor == True and global_monitor == None: global_monitor = self.procDirector.startMonitor( self.host, self.fuzzerData.port) print global_monitor time.sleep( 10 ) #clumsden added so pid_watcher has time to connect to monitor #! make it so logging message does not appear if reproducing (i.e. -r x-y cmdline arg is set) self.logger = None if not self.isReproduce: print "Logging to %s" % (self.outputDataFolderPath) self.logger = Logger(self.outputDataFolderPath) if self.args.dumpraw: if not isReproduce: self.DUMPDIR = self.outputDataFolderPath else: self.DUMPDIR = "dumpraw" try: os.mkdir("dumpraw") except: print "Unable to create dumpraw dir" pass self.exceptionProcessor = self.procDirector.exceptionProcessor() self.messageProcessor = self.procDirector.messageProcessor() ########## Begin fuzzing self.i = self.MIN_RUN_NUMBER - 1 if self.fuzzerData.shouldPerformTestRun else self.MIN_RUN_NUMBER self.failureCount = 0 self.loop_len = len(self.SEED_LOOP) # if --loop #will run one seed of the current instance of MutinyFuzzer def fuzz(self): args = self.args fuzzerData = self.fuzzerData host = self.host messageProcessor = self.messageProcessor retryRun = True while retryRun: retryRun = False lastMessageCollection = deepcopy(fuzzerData.messageCollection) wasCrashDetected = False print "\n** Sleeping for %.3f seconds **" % args.sleeptime time.sleep(args.sleeptime) try: try: print "\n\n%s: " % (self.fuzzerFilePath) if args.dumpraw: print "Performing single raw dump case: %d" % args.dumpraw self.performRun(fuzzerData, host, self.logger, messageProcessor, seed=args.dumpraw) elif self.i == self.MIN_RUN_NUMBER - 1: print "Performing test run without fuzzing..." self.performRun(fuzzerData, host, self.logger, messageProcessor, seed=-1) elif self.loop_len: print "Fuzzing with seed %d" % ( self.SEED_LOOP[self.i % loop_len]) self.performRun(fuzzerData, host, self.logger, messageProcessor, seed=self.SEED_LOOP[self.i % loop_len]) else: print "Fuzzing with seed %d" % (self.i) self.performRun(fuzzerData, host, self.logger, messageProcessor, seed=self.i) #if --quiet, (self.logger==None) => AttributeError if self.logAll: try: self.logger.outputLog(self.i, fuzzerData.messageCollection, "LogAll ") except AttributeError: pass #clumsden - so a crash detected by monitor is registered if global_monitor.crashEvent.isSet(): raise Exception() #clumsden end except Exception as e: if global_monitor.crashEvent.isSet(): print "Crash event detected" try: self.logger.outputLog(self.i, fuzzerData.messageCollection, "Crash event detected") exit( ) #clumsden - have this commented out if you don't want to stop after a crash is detected except AttributeError: pass global_monitor.crashEvent.clear() elif self.logAll: try: self.logger.outputLog(self.i, fuzzerData.messageCollection, "LogAll ") except AttributeError: pass if e.__class__ in MessageProcessorExceptions.all: # If it's a MessageProcessorException, assume the MP raised it during the run # Otherwise, let the MP know about the exception raise e else: self.exceptionProcessor.processException(e) # Will not get here if processException raises another exception print "Exception ignored: %s" % (str(e)) except LogCrashException as e: if self.failureCount == 0: try: print "MessageProcessor detected a crash" self.logger.outputLog(self.i, fuzzerData.messageCollection, str(e)) except AttributeError: pass if self.logAll: try: logger.outputLog(i, fuzzerData.messageCollection, "LogAll ") except AttributeError: pass self.failureCount = self.failureCount + 1 wasCrashDetected = True except AbortCurrentRunException as e: # Give up on the run early, but continue to the next test # This means the run didn't produce anything meaningful according to the processor print "Run aborted: %s" % (str(e)) except RetryCurrentRunException as e: # Same as AbortCurrentRun but retry the current test rather than skipping to next print "Retrying current run: %s" % (str(e)) # Slightly sketchy - a continue *should* just go to the top of the while without changing i retryRun = True continue except LogAndHaltException as e: if self.logger: self.logger.outputLog(self.i, fuzzerData.messageCollection, str(e)) print "Received LogAndHaltException, logging and halting" else: print "Received LogAndHaltException, halting but not logging (quiet mode)" exit() except LogLastAndHaltException as e: if self.logger: if self.i > self.MIN_RUN_NUMBER: print "Received LogLastAndHaltException, logging last run and halting" if self.MIN_RUN_NUMBER == self.MAX_RUN_NUMBER: #in case only 1 case is run self.logger.outputLastLog(self.i, lastMessageCollection, str(e)) print "Logged case %d" % self.i else: self.logger.outputLastLog(self.i - 1, lastMessageCollection, str(e)) else: print "Received LogLastAndHaltException, skipping logging (due to last run being a test run) and halting" else: print "Received LogLastAndHaltException, halting but not logging (quiet mode)" exit() except HaltException as e: print "Received HaltException halting" exit() if wasCrashDetected: if self.failureCount < fuzzerData.failureThreshold: print "Failure %d of %d allowed for seed %d" % ( self.failureCount, fuzzerData.failureThreshold, self.i) print "The test run didn't complete, continuing after %d seconds..." % ( fuzzerData.failureTimeout) time.sleep(fuzzerData.failureTimeout) else: print "Failed %d times, moving to next test." % ( self.failureCount) self.failureCount = 0 self.i += 1 else: self.i += 1 # Stop if we have a maximum and have hit it if self.MAX_RUN_NUMBER >= 0 and self.i > self.MAX_RUN_NUMBER: exit() if args.dumpraw: exit() # Perform a fuzz run. # If seed is -1, don't perform fuzzing (test run) def performRun(self, fuzzerData, host, logger, messageProcessor, seed=-1): # Before doing anything, set up logger # Otherwise, if connection is refused, we'll log last, but it will be wrong if logger != None: logger.resetForNewRun() # We don't perform DNS resolution, but always automatically type "localhost" # ... really need to go ahead and add DNS resolution soon if host == "localhost": host = "127.0.0.1" # cheap testing for ipv6/ipv4/unix # don't think it's worth using regex for this, since the user # will have to actively go out of their way to subvert this. if "." in host: socket_family = socket.AF_INET addr = (host, fuzzerData.port) elif ":" in host: socket_family = socket.AF_INET6 addr = (host, fuzzerData.port) else: socket_family = socket.AF_UNIX addr = (host) #just in case filename is like "./asdf" !=> AF_INET if "/" in host: socket_family = socket.AF_UNIX addr = (host) # Call messageprocessor preconnect callback if it exists try: messageProcessor.preConnect(seed, host, fuzzerData.port) except AttributeError: pass # for TCP/UDP/RAW support if fuzzerData.proto == "tcp": connection = socket.socket(socket_family, socket.SOCK_STREAM) # Don't connect yet, until after we do any binding below elif fuzzerData.proto == "tls": try: _create_unverified_https_context = ssl._create_unverified_context except AttributeError: # Legacy Python that doesn't verify HTTPS certificates by default pass else: # Handle target environment that doesn't support HTTPS verification ssl._create_default_https_context = _create_unverified_https_context tcpConnection = socket.socket(socket_family, socket.SOCK_STREAM) connection = ssl.wrap_socket(tcpConnection) # Don't connect yet, until after we do any binding below elif fuzzerData.proto == "udp": connection = socket.socket(socket_family, socket.SOCK_DGRAM) # PROTO = dictionary of assorted L3 proto => proto number # e.g. "icmp" => 1 elif fuzzerData.proto in PROTO: connection = socket.socket(socket_family, socket.SOCK_RAW, PROTO[fuzzerData.proto]) if fuzzerData.proto != "raw": connection.setsockopt(socket.IPPROTO_IP, socket.IP_HDRINCL, 0) addr = (host, 0) try: connection = socket.socket(socket_family, socket.SOCK_RAW, PROTO[fuzzerData.proto]) except Exception as e: print e print "Unable to create raw socket, please verify that you have sudo access" sys.exit(0) elif fuzzerData.proto == "L2raw": connection = socket.socket(socket.AF_PACKET, socket.SOCK_RAW, 0x0300) connection.bind( ('ens160', 0)) #clumsden TODO replace hardcoded iface for layer2 traffic #clumsden TODO - experimental branch has addr=(host,0) else: addr = (host, 0) try: #test if it's a valid number connection = socket.socket(socket_family, socket.SOCK_RAW, int(fuzzerData.proto)) connection.setsockopt(socket.IPPROTO_IP, socket.IP_HDRINCL, 0) except Exception as e: print e print "Unable to create raw socket, please verify that you have sudo access" sys.exit(0) if fuzzerData.proto == "tcp" or fuzzerData.proto == "udp" or fuzzerData.proto == "tls": # Specifying source port or address is only supported for tcp and udp currently if fuzzerData.sourcePort != -1: # Only support right now for tcp or udp, but bind source port address to something # specific if requested if fuzzerData.sourceIP != "" or fuzzerData.sourceIP != "0.0.0.0": connection.bind( (fuzzerData.sourceIP, fuzzerData.sourcePort)) else: # User only specified a port, not an IP connection.bind(('0.0.0.0', fuzzerData.sourcePort)) elif fuzzerData.sourceIP != "" and fuzzerData.sourceIP != "0.0.0.0": # No port was specified, so 0 should auto-select connection.bind((fuzzerData.sourceIP, 0)) if fuzzerData.proto == "tcp" or fuzzerData.proto == "tls": # Now that we've had a chance to bind as necessary, connect connection.connect(addr) i = 0 for i in range(0, len(fuzzerData.messageCollection.messages)): message = fuzzerData.messageCollection.messages[i] # Go ahead and revert any fuzzing or messageprocessor changes before proceeding message.resetAlteredMessage() if message.isOutbound(): # Primarily used for deciding how to handle preFuzz/preSend callbacks doesMessageHaveSubcomponents = len(message.subcomponents) > 1 # Get original subcomponents for outbound callback only once originalSubcomponents = map( lambda subcomponent: subcomponent.getOriginalByteArray(), message.subcomponents) if doesMessageHaveSubcomponents: # For message with subcomponents, call prefuzz on fuzzed subcomponents for j in range(0, len(message.subcomponents)): subcomponent = message.subcomponents[j] # Note: we WANT to fetch subcomponents every time on purpose # This way, if user alters subcomponent[0], it's reflected when # we call the function for subcomponent[1], etc actualSubcomponents = map( lambda subcomponent: subcomponent. getAlteredByteArray(), message.subcomponents) prefuzz = messageProcessor.preFuzzSubcomponentProcess( subcomponent.getAlteredByteArray(), MessageProcessorExtraParams( i, j, subcomponent.isFuzzed, originalSubcomponents, actualSubcomponents)) subcomponent.setAlteredByteArray(prefuzz) else: # If no subcomponents, call prefuzz on ENTIRE message actualSubcomponents = map( lambda subcomponent: subcomponent.getAlteredByteArray( ), message.subcomponents) prefuzz = messageProcessor.preFuzzProcess( actualSubcomponents[0], MessageProcessorExtraParams(i, -1, message.isFuzzed, originalSubcomponents, actualSubcomponents)) message.subcomponents[0].setAlteredByteArray(prefuzz) # Skip fuzzing for seed == -1 if seed > -1: # Now run the fuzzer for each fuzzed subcomponent for subcomponent in message.subcomponents: if subcomponent.isFuzzed: radamsa = subprocess.Popen( [RADAMSA, "--seed", str(seed)], stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE) byteArray = subcomponent.getAlteredByteArray() (fuzzedByteArray, error_output) = radamsa.communicate( input=byteArray) fuzzedByteArray = bytearray(fuzzedByteArray) subcomponent.setAlteredByteArray(fuzzedByteArray) # Fuzzing has now been done if this message is fuzzed # Always call preSend() regardless for subcomponents if there are any if doesMessageHaveSubcomponents: for j in range(0, len(message.subcomponents)): subcomponent = message.subcomponents[j] # See preFuzz above - we ALWAYS regather this to catch any updates between # callbacks from the user actualSubcomponents = map( lambda subcomponent: subcomponent. getAlteredByteArray(), message.subcomponents) presend = messageProcessor.preSendSubcomponentProcess( subcomponent.getAlteredByteArray(), MessageProcessorExtraParams( i, j, subcomponent.isFuzzed, originalSubcomponents, actualSubcomponents)) subcomponent.setAlteredByteArray(presend) # Always let the user make any final modifications pre-send, fuzzed or not actualSubcomponents = map( lambda subcomponent: subcomponent.getAlteredByteArray(), message.subcomponents) byteArrayToSend = messageProcessor.preSendProcess( message.getAlteredMessage(), MessageProcessorExtraParams(i, -1, message.isFuzzed, originalSubcomponents, actualSubcomponents)) if self.args.dumpraw: loc = os.path.join( DUMPDIR, "%d-outbound-seed-%d" % (i, self.args.dumpraw)) if message.isFuzzed: loc += "-fuzzed" with open(loc, "wb") as f: f.write(repr(str(byteArrayToSend))[1:-1]) self.sendPacket(connection, addr, byteArrayToSend) else: # Receiving packet from server messageByteArray = message.getAlteredMessage() data = self.receivePacket(connection, addr, len(messageByteArray)) if data == messageByteArray: print "\tReceived expected response" if logger != None: logger.setReceivedMessageData(i, data) messageProcessor.postReceiveProcess( data, MessageProcessorExtraParams(i, -1, False, [messageByteArray], [data])) if self.args.dumpraw: loc = os.path.join( DUMPDIR, "%d-inbound-seed-%d" % (i, self.args.dumpraw)) with open(loc, "wb") as f: f.write(repr(str(data))[1:-1]) if logger != None: logger.setHighestMessageNumber(i) i += 1 connection.close() def receivePacket(self, connection, addr, bytesToRead): readBufSize = 4096 connection.settimeout(self.fuzzerData.receiveTimeout) if connection.type == socket.SOCK_STREAM or connection.type == socket.SOCK_DGRAM: response = bytearray(connection.recv(readBufSize)) else: response = bytearray(connection.recvfrom(readBufSize, addr)) if len(response) == 0: # If 0 bytes are recv'd, the server has closed the connection # per python documentation raise ConnectionClosedException("Server has closed the connection") if bytesToRead > readBufSize: # If we're trying to read > 4096, don't actually bother trying to guarantee we'll read 4096 # Just keep reading in 4096 chunks until we should have read enough, and then return # whether or not it's as much data as expected i = readBufSize while i < bytesToRead: response += bytearray(connection.recv(readBufSize)) i += readBufSize print "\tReceived %d bytes" % (len(response)) if DEBUG_MODE: print "\tReceived: %s" % (response) return response # Takes a socket and outbound data packet (byteArray), sends it out. # If debug mode is enabled, we print out the raw bytes def sendPacket(self, connection, addr, outPacketData): connection.settimeout(self.fuzzerData.receiveTimeout) if connection.type == socket.SOCK_STREAM: connection.send(outPacketData) elif connection.type == socket.SOCK_RAW: #for L2raw clumsden start connection.send( outPacketData ) # clumsden, ran into trouble with this later and needed sendto(outPacketData,addr)...what it was orginally? else: connection.sendto(outPacketData, addr) print "\tSent %d byte packet" % (len(outPacketData)) if DEBUG_MODE: print "\tSent: %s" % (outPacketData) print "\tRaw Bytes: %s" % ( Message.serializeByteArray(outPacketData))