def __init__(self, port, id_, version):
self._id = id_
self._version = version
self._server = KRPCServer(port, self._version)
self._rt = PrefixRoutingTable()
# Thread details
self._shutdown_flag = False
self._thread = None
# default handler
self.handler = self.default_handler
# Behaviour configuration
# Am I actively seeking out other nodes?
self.active_discovery = True
# After how many seconds should i do another self-lookup?
self.self_find_delay = 180.0
# How many active node discovery attempts between self-lookups?
self.active_discoveries = 10
# Session key
self._key = os.urandom(20) # 20 random bytes == 160 bits
def __init__(self, port, id_, version):
self._id = id_
self._version = version
self._server = KRPCServer(port, self._version)
self._rt = PrefixRoutingTable()
# Thread details
self._shutdown_flag = False
self._thread = None
# default handler
self.handler = self.default_handler
# Behaviour configuration
# Am I actively seeking out other nodes?
self.active_discovery = True
# After how many seconds should i do another self-lookup?
self.self_find_delay = 180.0
# How many active node discovery attempts between self-lookups?
self.active_discoveries = 10
# Session key
self._key = os.urandom(20) # 20 random bytes == 160 bits
class DHT(object):
def __init__(self, port, id_, version):
self._id = id_
self._version = version
self._server = KRPCServer(port, self._version)
self._rt = PrefixRoutingTable()
# Thread details
self._shutdown_flag = False
self._thread = None
# default handler
self.handler = self.default_handler
# Behaviour configuration
# Am I actively seeking out other nodes?
self.active_discovery = True
# After how many seconds should i do another self-lookup?
self.self_find_delay = 180.0
# How many active node discovery attempts between self-lookups?
self.active_discoveries = 10
# Session key
self._key = os.urandom(20) # 20 random bytes == 160 bits
def _get_id(self, target):
# Retrieve ID to use to communicate with target node
return self._id
def start(self):
"""
Start the DHT node
"""
self._server.start()
self._server.handler = self.handler
# Add the default nodes
DEFAULT_CONNECT_INFO = (
socket.gethostbyaddr("router.bittorrent.com")[2][0], 6881)
DEFAULT_NODE = Node(DEFAULT_CONNECT_INFO)
DEFAULT_ID = self._server.ping(os.urandom(20), DEFAULT_NODE)['id']
self._rt.update_entry(DEFAULT_ID, DEFAULT_NODE)
# Start our event thread
self._thread = threading.Thread(target=self._pump)
self._thread.daemon = True
self._thread.start()
def shutdown(self):
self._server.shutdown()
def __enter__(self):
self.start()
def __exit__(self, type_, value, traceback):
self.shutdown()
def _pump(self):
"""
Thread that maintains DHT connectivity and does
routing table housekeeping.
Started by self.start()
The very first thing this function does, is look up itself
in the DHT. This connects it to neighbouring nodes and enables
it to give reasonable answers to incoming queries.
Afterward we look up random nodes to increase our connectedness
and gather information about the DHT as a whole
"""
# Try to establish links to close nodes
logger.info("Establishing connections to DHT")
self.find_node(self._id)
delay = self.self_find_delay
if self.active_discovery:
delay /= (self.active_discoveries + 1)
iteration = 0
while True:
try:
time.sleep(delay)
iteration += 1
if self.active_discovery and iteration % (
self.active_discoveries + 1) != 0:
target = hashlib.sha1("this is my salt 2348724" +
str(iteration) + self._id).digest()
self.find_node(target)
logger.info(
"Tracing done, routing table contains %d nodes",
self._rt.node_count())
else:
# Regular maintenance:
# Find N random nodes. Execute a find_node() on them.
# toss them if they come up empty.
n = self._rt.sample(self._id, 10, 1)
for node_id, c in n:
try:
r = self._server.find_node(self._id, c, self._id)
if "nodes" in r:
self._process_incoming_nodes(r["nodes"])
except KRPCTimeout:
# The node did not reply.
# Blacklist it.
self._rt.bad_node(node_id, c)
logger.info("Cleanup, routing table contains %d nodes",
self._rt.node_count())
except:
# This loop should run forever. If we get into trouble, log
# the exception and carry on.
logger.critical("Exception in DHT maintenance thread:\n\n" +
traceback.format_exc())
def _process_incoming_nodes(self, bnodes):
# Add them to the routing table
for node_id, node_c in decode_nodes(bnodes):
self._rt.update_entry(node_id, Node(node_c))
def _recurse(self, target, function, max_attempts=10, result_key=None):
"""
Recursively query the DHT, following "nodes" replies
until we hit the desired key
This is the workhorse function used by all recursive queries.
"""
logger.debug("Recursing to target %r" % target.encode("hex"))
attempts = 0
while attempts < max_attempts:
for id_, node in self._rt.get_close_nodes(target):
try:
r = function(self._get_id(id_), node, target)
logger.debug("Recursion results from %r ", node.c)
attempts += 1
if result_key and result_key in r:
return r[result_key]
if "nodes" in r:
self._process_incoming_nodes(r["nodes"])
except KRPCTimeout:
# The node did not reply.
# Blacklist it.
self._rt.bad_node(id_, node)
except KRPCError:
# Sometimes we just flake out due to UDP being unreliable
# Don't sweat it, just log and carry on.
logger.error("KRPC Error:\n\n" + traceback.format_exc())
if result_key:
# We were expecting a result, but we did not find it!
# Raise the NotFoundError exception instead of returning None
raise NotFoundError
def find_node(self, target, attempts=10):
"""
Recursively call the find_node function to get as
close as possible to the target node
"""
logger.debug("Tracing to %r" % target.encode("hex"))
self._recurse(target, self._server.find_node, max_attempts=attempts)
def get_peers(self, info_hash, attempts=10):
"""
Recursively call the get_peers function to fidn peers
for the given info_hash
"""
logger.debug("Finding peers for %r" % info_hash.encode("hex"))
return self._recurse(info_hash,
self._server.get_peers,
result_key="values",
max_attempts=attempts)
def default_handler(self, rec, c):
"""
Process incoming requests
"""
logger.info("REQUEST: %r %r" % (c, rec))
# Use the request to update the routing table
peer_id = rec["a"]["id"]
self._rt.update_entry(peer_id, Node(c))
# Skeleton response
resp = {
"y": "r",
"t": rec["t"],
"r": {
"id": self._get_id(peer_id)
},
"v": self._version
}
if rec["q"] == "ping":
self._server.send_krpc_reply(resp, c)
elif rec["q"] == "find_node":
target = rec["a"]["target"]
resp["r"]["id"] = self._get_id(target)
resp["r"]["nodes"] = encode_nodes(self._rt.get_close_nodes(target))
self._server.send_krpc_reply(resp, c)
elif rec["q"] == "get_peers":
# Provide a token so we can receive announces
# The token is generated using HMAC and a secret
# session key, so we don't have to remember it.
# Token is based on nodes id, connection details
# torrent infohash to avoid clashes in NAT scenarios.
info_hash = rec["a"]["info_hash"]
resp["r"]["id"] = self._get_id(info_hash)
token = hmac.new(self._key, info_hash + peer_id + str(c),
hashlib.sha1).digest()
resp["r"]["token"] = token
# We don't actually keep any peer administration, so we
# always send back the closest nodes
resp["r"]["nodes"] = encode_nodes(
self._rt.get_close_nodes(info_hash))
self._server.send_krpc_reply(resp, c)
elif rec["q"] == "announce_peer":
# First things first, validate the token.
info_hash = rec["a"]["info_hash"]
resp["r"]["id"] = self._get_id(info_hash)
peer_id = rec["a"]["id"]
token = hmac.new(self._key, info_hash + peer_id + str(c),
hashlib.sha1).digest()
if token != rec["a"]["token"]:
return # Ignore the request
else:
# We don't actually keep any peer administration, so we
# just acknowledge.
self._server.send_krpc_reply(resp, c)
else:
logger.error("Unknown request in query %r" % rec)
class DHT(object):
def __init__(self, port, id_, version):
self._id = id_
self._version = version
self._server = KRPCServer(port, self._version)
self._rt = PrefixRoutingTable()
# Thread details
self._shutdown_flag = False
self._thread = None
# default handler
self.handler = self.default_handler
# Behaviour configuration
# Am I actively seeking out other nodes?
self.active_discovery = True
# After how many seconds should i do another self-lookup?
self.self_find_delay = 180.0
# How many active node discovery attempts between self-lookups?
self.active_discoveries = 10
# Session key
self._key = os.urandom(20) # 20 random bytes == 160 bits
def _get_id(self, target):
# Retrieve ID to use to communicate with target node
return self._id
def start(self):
"""
Start the DHT node
"""
self._server.start()
self._server.handler = self.handler
# Add the default nodes
DEFAULT_CONNECT_INFO = ('67.215.242.139', 6881) #(socket.gethostbyaddr("router.bittorrent.com")[2][0], 6881)
DEFAULT_NODE = Node(DEFAULT_CONNECT_INFO)
DEFAULT_ID = self._server.ping(os.urandom(20), DEFAULT_NODE)['id']
self._rt.update_entry(DEFAULT_ID, DEFAULT_NODE)
# Start our event thread
self._thread = threading.Thread(target=self._pump)
self._thread.daemon = True
self._thread.start()
def shutdown(self):
self._server.shutdown()
def __enter__(self):
self.start()
def __exit__(self, type_, value, traceback):
self.shutdown()
def _pump(self):
"""
Thread that maintains DHT connectivity and does
routing table housekeeping.
Started by self.start()
The very first thing this function does, is look up itself
in the DHT. This connects it to neighbouring nodes and enables
it to give reasonable answers to incoming queries.
Afterward we look up random nodes to increase our connectedness
and gather information about the DHT as a whole
"""
# Try to establish links to close nodes
logger.info("Establishing connections to DHT")
self.find_node(self._id)
delay = self.self_find_delay
if self.active_discovery:
delay /= (self.active_discoveries + 1)
iteration = 0
while True:
try:
time.sleep(delay)
iteration += 1
if self.active_discovery and iteration % (self.active_discoveries + 1) != 0:
target = hashlib.sha1("this is my salt 2348724" + str(iteration) + self._id).digest()
self.find_node(target)
logger.info("Tracing done, routing table contains %d nodes", self._rt.node_count())
else:
# Regular maintenance:
# Find N random nodes. Execute a find_node() on them.
# toss them if they come up empty.
n = self._rt.sample(self._id, 10, 1)
for node_id, c in n:
try:
r = self._server.find_node(self._id, c, self._id)
if "nodes" in r:
self._process_incoming_nodes(r["nodes"])
except KRPCTimeout:
# The node did not reply.
# Blacklist it.
self._rt.bad_node(node_id, c)
logger.info("Cleanup, routing table contains %d nodes", self._rt.node_count())
except:
# This loop should run forever. If we get into trouble, log
# the exception and carry on.
logger.critical("Exception in DHT maintenance thread:\n\n" + traceback.format_exc())
def _process_incoming_nodes(self, bnodes):
# Add them to the routing table
for node_id, node_c in decode_nodes(bnodes):
self._rt.update_entry(node_id, Node(node_c))
def _recurse(self, target, function, max_attempts=10, result_key=None):
"""
Recursively query the DHT, following "nodes" replies
until we hit the desired key
This is the workhorse function used by all recursive queries.
"""
logger.debug("Recursing to target %r" % target.encode("hex"))
attempts = 0
while attempts < max_attempts:
for id_, node in self._rt.get_close_nodes(target):
try:
r = function(self._get_id(id_), node, target)
logger.debug("Recursion results from %r ", node.c)
attempts += 1
if result_key and result_key in r:
return r[result_key]
if "nodes" in r:
self._process_incoming_nodes(r["nodes"])
except KRPCTimeout:
# The node did not reply.
# Blacklist it.
self._rt.bad_node(id_, node)
except KRPCError:
# Sometimes we just flake out due to UDP being unreliable
# Don't sweat it, just log and carry on.
logger.error("KRPC Error:\n\n" + traceback.format_exc())
if result_key:
# We were expecting a result, but we did not find it!
# Raise the NotFoundError exception instead of returning None
raise NotFoundError
def find_node(self, target, attempts=10):
"""
Recursively call the find_node function to get as
close as possible to the target node
"""
logger.debug("Tracing to %r" % target.encode("hex"))
self._recurse(target, self._server.find_node, max_attempts=attempts)
def get_peers(self, info_hash, attempts=10):
"""
Recursively call the get_peers function to fidn peers
for the given info_hash
"""
logger.debug("Finding peers for %r" % info_hash.encode("hex"))
return self._recurse(info_hash, self._server.get_peers, result_key="values", max_attempts=attempts)
def default_handler(self, rec, c):
"""
Process incoming requests
"""
logger.info("REQUEST: %r %r" % (c, rec))
# Use the request to update the routing table
peer_id = rec["a"]["id"]
#print peer_id.encode('base64'), self._get_id(peer_id).encode('base64')
if self._get_id(peer_id) == peer_id:
# don't talk to yourself.
return
self._rt.update_entry(peer_id, Node(c))
# Skeleton response
resp = {"y": "r", "t": rec["t"], "r": {"id": self._get_id(peer_id)}, "v": self._version}
if rec["q"] == "ping":
self._server.send_krpc_reply(resp, c)
elif rec["q"] == "find_node":
target = rec["a"]["target"]
resp["r"]["id"] = self._get_id(target)
resp["r"]["nodes"] = encode_nodes(self._rt.get_close_nodes(target))
self._server.send_krpc_reply(resp, c)
elif rec["q"] == "get_peers":
# Provide a token so we can receive announces
# The token is generated using HMAC and a secret
# session key, so we don't have to remember it.
# Token is based on nodes id, connection details
# torrent infohash to avoid clashes in NAT scenarios.
info_hash = rec["a"]["info_hash"]
resp["r"]["id"] = self._get_id(info_hash)
token = hmac.new(self._key, info_hash + peer_id + str(c), hashlib.sha1).digest()
resp["r"]["token"] = token
# We don't actually keep any peer administration, so we
# always send back the closest nodes
resp["r"]["nodes"] = encode_nodes(self._rt.get_close_nodes(info_hash))
self._server.send_krpc_reply(resp, c)
elif rec["q"] == "announce_peer":
# First things first, validate the token.
info_hash = rec["a"]["info_hash"]
resp["r"]["id"] = self._get_id(info_hash)
peer_id = rec["a"]["id"]
token = hmac.new(self._key, info_hash + peer_id + str(c), hashlib.sha1).digest()
if token != rec["a"]["token"]:
return # Ignore the request
else:
# We don't actually keep any peer administration, so we
# just acknowledge.
self._server.send_krpc_reply(resp, c)
else:
logger.error("Unknown request in query %r" % rec)
class DHT(object):
def __init__(self, port, id_, version):
self._id = id_
self._version = version
self._server = KRPCServer(port, self._version)
self._rt = PrefixRoutingTable()
# Thread details
self._shutdown_flag = False
self._thread = None
# default handler
self.handler = self.default_handler
# Behaviour configuration
# Am I actively seeking out other nodes?
self.active_discovery = True
# After how many seconds should i do another self-lookup?
self.self_find_delay = 180.0
# How many active node discovery attempts between self-lookups?
self.active_discoveries = 10
# Session key
self._key = os.urandom(20) # 20 random bytes == 160 bits
#print("Finished __init__.")
def _get_id(self, target):
# Retrieve ID to use to communicate with target node
return self._id
def start(self):
"""
Start the DHT node
"""
#print("In start.")
self._server.start()
self._server.handler = self.handler
# Add the default nodes
# socket.gethostbyaddr returns (hostname, aliaslist, ipaddrlist)
# So, this uses the first alternate ip address of router.bittorrent.com
# according to DNS resolution.
AltIPs = socket.gethostbyaddr("router.bittorrent.com")[2]
# DEFAULT_CONNECT_INFO = (random.choice(AltIPs), 6881)
# # Default_Node is assigned a tuple of (ip, port)
# DEFAULT_NODE = Node(DEFAULT_CONNECT_INFO)
# DEFAULT_ID = self._server.ping(os.urandom(20), DEFAULT_NODE)['id']
# self._rt.update_entry(DEFAULT_ID, DEFAULT_NODE)
# Prior behaviour was to pick one of the router's alt-ips, this adds
# all of them..
for ip in AltIPs:
Connect_Info = (ip, 6881)
# Default_Node is assigned a tuple of (ip, port)
IP_Node = Node(Connect_Info)
IP_ID = self._server.ping(os.urandom(20), IP_Node)['id']
logger.info("Adding bootstrap alt-ip from router.bittorrent.com: IP: {0}, NodeID: {1}".format(IP_Node, IP_ID))
self._rt.update_entry(IP_ID, IP_Node)
# Start our event thread
self._thread = threading.Thread(target=self._pump)
self._thread.daemon = True
self._thread.start()
#print("Finished start.")
def shutdown(self):
self._server.shutdown()
def __enter__(self):
#print("In __enter__.")
self.start()
def __exit__(self, type_, value, traceback):
self.shutdown()
def _pump(self):
"""
Thread that maintains DHT connectivity and does
routing table housekeeping.
Started by self.start()
The very first thing this function does, is look up itself
in the DHT. This connects it to neighbouring nodes and enables
it to give reasonable answers to incoming queries.
Afterward we look up random nodes to increase our connectedness
and gather information about the DHT as a whole
"""
#print("Started _pump thread.")
# Try to establish links to close nodes
logger.info("Establishing connections to DHT")
found_self = False
while not found_self:
try:
self.find_node(self._id)
found_self = True
except:
logger.critical("Exception while starting DHT Maintainence thread:\n\n" + traceback.format_exc())
time.sleep(1)
delay = self.self_find_delay
if self.active_discovery:
delay //= (self.active_discoveries + 1)
logger.info("Finished establishing connections to DHT, beginning maintenance.")
iteration = 0
while True:
try:
time.sleep(delay)
iteration += 1
if self.active_discovery and iteration % (self.active_discoveries + 1) != 0:
target = hashlib.sha1("this is my salt 2348724" + str(iteration) + self._id).digest()
self.find_node(target)
logger.info("Tracing done, routing table contains %d nodes", self._rt.node_count())
else:
# Regular maintenance:
# Find N random nodes. Execute a find_node() on them.
# toss them if they come up empty.
n = self._rt.sample(self._id, 10, 1)
for node_id, c in n:
try:
#print("In _pump: calling self._server.find_node()")
r = self._server.find_node(self._id, c, self._id)
#print("In _pump: finished self._server.find_node()")
if "nodes" in r:
self._process_incoming_nodes(r["nodes"])
except KRPCTimeout:
# The node did not reply.
# Blacklist it.
self._rt.bad_node(node_id, c)
logger.info("Cleanup, routing table contains {0} nodes".format(self._rt.node_count()))
except:
# This loop should run forever. If we get into trouble, log
# the exception and carry on.
logger.critical("Exception in DHT maintenance thread:\n\n" + traceback.format_exc())
def _process_incoming_nodes(self, bnodes):
# Add them to the routing table
for node_id, node_c in decode_nodes(bnodes):
self._rt.update_entry(node_id, Node(node_c))
def _recurse(self, target, function, max_attempts=10, result_key=None):
"""
Recursively query the DHT, following "nodes" replies
until we hit the desired key
This is the workhorse function used by all recursive queries.
"""
#print("In _recurse.")
if isinstance(target, bytes):
target_hex = binascii.hexlify(target).decode()
else:
target_hex = target
logger.debug("Recursing to target {0}".format(target))
attempts = 0
while attempts < max_attempts:
close_nodes = self._rt.get_close_nodes(target)
if not close_nodes:
raise NotFoundError("No close nodes found with self.rt.get_close_nodes for "+str(target)+\
" Current routing table: "+str(self._rt._nodes))
for id_, node in close_nodes:
try:
#print("Calling function", function, "in _recurse.")
r = function(self._get_id(id_), node, target)
#print("Finished calling function in _recurse.")
logger.debug("Recursion results from %r ", node.c)
attempts += 1
if result_key and result_key in r:
return r[result_key]
if "nodes" in r:
self._process_incoming_nodes(r["nodes"])
except KRPCTimeout:
# The node did not reply.
# Blacklist it.
if self._rt.node_count() > 8:
logger.error("Node timed out: blacklisting {0}".format(node.c))
self._rt.bad_node(id_, node)
else:
logger.error("Node timed out: Would blacklist, but only 8 nodes known. Node: {0}".format(node.c))
continue
except KRPCError:
# Sometimes we just flake out due to UDP being unreliable
# Don't sweat it, just log and carry on.
logger.error("KRPC Error:\n\n" + traceback.format_exc())
if result_key:
# We were expecting a result, but we did not find it!
# Raise the NotFoundError exception instead of returning None
raise NotFoundError
#print("Finished _recurse.")
def find_node(self, target, attempts=10):
"""
Recursively call the find_node function to get as
close as possible to the target node
"""
if isinstance(target, bytes):
target_hex = binascii.hexlify(target).decode()
else:
target_hex = target
logger.debug("Tracing to {0}".format(target_hex))
self._recurse(target, self._server.find_node, max_attempts=attempts)
def get_peers(self, info_hash, attempts=10):
"""
Recursively call the get_peers function to fidn peers
for the given info_hash
"""
if isinstance(info_hash, bytes):
info_hash_hex = binascii.hexlify(info_hash).decode()
else:
info_hash_hex = info_hash
logger.debug("Finding peers for {0}".format(info_hash_hex))
return self._recurse(info_hash, self._server.get_peers, result_key="values", max_attempts=attempts)
def default_handler(self, rec, c):
"""
Process incoming requests
"""
logger.info("REQUEST: %r %r" % (c, rec))
# Use the request to update the routing table
peer_id = rec["a"]["id"]
self._rt.update_entry(peer_id, Node(c))
# Skeleton response
resp = {"y": "r", "t": rec["t"], "r": {"id": self._get_id(peer_id)}, "v": self._version}
if rec["q"] == b"ping":
self._server.send_krpc_reply(resp, c)
elif rec["q"] == b"find_node":
target = rec["a"]["target"]
resp["r"]["id"] = self._get_id(target)
resp["r"]["nodes"] = encode_nodes(self._rt.get_close_nodes(target))
self._server.send_krpc_reply(resp, c)
elif rec["q"] == b"get_peers":
# Provide a token so we can receive announces
# The token is generated using HMAC and a secret
# session key, so we don't have to remember it.
# Token is based on nodes id, connection details
# torrent infohash to avoid clashes in NAT scenarios.
info_hash = rec["a"]["info_hash"]
resp["r"]["id"] = self._get_id(info_hash)
token = hmac.new(self._key, info_hash + peer_id + str(c), hashlib.sha1).digest()
resp["r"]["token"] = token
# We don't actually keep any peer administration, so we
# always send back the closest nodes
resp["r"]["nodes"] = encode_nodes(self._rt.get_close_nodes(info_hash))
self._server.send_krpc_reply(resp, c)
elif rec["q"] == b"announce_peer":
# First things first, validate the token.
info_hash = rec["a"]["info_hash"]
resp["r"]["id"] = self._get_id(info_hash)
peer_id = rec["a"]["id"]
token = hmac.new(self._key, info_hash + peer_id + str(c), hashlib.sha1).digest()
if token != rec["a"]["token"]:
return # Ignore the request
else:
# We don't actually keep any peer administration, so we
# just acknowledge.
self._server.send_krpc_reply(resp, c)
else:
logger.error("Unknown request in query %r" % rec)