#!/usr/bin/env python3 # Copyright (c) 2014-2016 The Bitcoin Core developers # Distributed under the MIT software license, see the accompanying # file COPYING or http://www.opensource.org/licenses/mit-license.php. """Test behavior of headers messages to announce blocks. Setup: - Two nodes, two p2p connections to node0. One p2p connection should only ever receive inv's (omitted from testing description below, this is our control). Second node is used for creating reorgs. Part 1: No headers announcements before "sendheaders" a. node mines a block [expect: inv] send getdata for the block [expect: block] b. node mines another block [expect: inv] send getheaders and getdata [expect: headers, then block] c. node mines another block [expect: inv] peer mines a block, announces with header [expect: getdata] d. node mines another block [expect: inv] Part 2: After "sendheaders", headers announcements should generally work. a. peer sends sendheaders [expect: no response] peer sends getheaders with current tip [expect: no response] b. node mines a block [expect: tip header] c. for N in 1, ..., 10: * for announce-type in {inv, header} - peer mines N blocks, announces with announce-type [ expect: getheaders/getdata or getdata, deliver block(s) ] - node mines a block [ expect: 1 header ] Part 3: Headers announcements stop after large reorg and resume after getheaders or inv from peer. - For response-type in {inv, getheaders} * node mines a 7 block reorg [ expect: headers announcement of 8 blocks ] * node mines an 8-block reorg [ expect: inv at tip ] * peer responds with getblocks/getdata [expect: inv, blocks ] * node mines another block [ expect: inv at tip, peer sends getdata, expect: block ] * node mines another block at tip [ expect: inv ] * peer responds with getheaders with an old hashstop more than 8 blocks back [expect: headers] * peer requests block [ expect: block ] * node mines another block at tip [ expect: inv, peer sends getdata, expect: block ] * peer sends response-type [expect headers if getheaders, getheaders/getdata if mining new block] * node mines 1 block [expect: 1 header, peer responds with getdata] Part 4: Test direct fetch behavior a. Announce 2 old block headers. Expect: no getdata requests. b. Announce 3 new blocks via 1 headers message. Expect: one getdata request for all 3 blocks. (Send blocks.) c. Announce 1 header that forks off the last two blocks. Expect: no response. d. Announce 1 more header that builds on that fork. Expect: one getdata request for two blocks. e. Announce 16 more headers that build on that fork. Expect: getdata request for 14 more blocks. f. Announce 1 more header that builds on that fork. Expect: no response. Part 5: Test handling of headers that don't connect. a. Repeat 10 times: 1. Announce a header that doesn't connect. Expect: getheaders message 2. Send headers chain. Expect: getdata for the missing blocks, tip update. b. Then send 9 more headers that don't connect. Expect: getheaders message each time. c. Announce a header that does connect. Expect: no response. d. Announce 49 headers that don't connect. Expect: getheaders message each time. e. Announce one more that doesn't connect. Expect: disconnect. """ from test_framework.mininode import * from test_framework.test_framework import BitcoinTestFramework from test_framework.util import * from test_framework.blocktools import create_block, create_coinbase direct_fetch_response_time = 0.05 class TestNode(NodeConnCB): def __init__(self): super().__init__() self.block_announced = False self.last_blockhash_announced = None def clear_last_announcement(self): with mininode_lock: self.block_announced = False self.last_message.pop("inv", None) self.last_message.pop("headers", None) # Request data for a list of block hashes def get_data(self, block_hashes): msg = msg_getdata() for x in block_hashes: msg.inv.append(CInv(2, x)) self.connection.send_message(msg) def get_headers(self, locator, hashstop): msg = msg_getheaders() msg.locator.vHave = locator msg.hashstop = hashstop self.connection.send_message(msg) def send_block_inv(self, blockhash): msg = msg_inv() msg.inv = [CInv(2, blockhash)] self.connection.send_message(msg) def on_inv(self, conn, message): self.block_announced = True self.last_blockhash_announced = message.inv[-1].hash def on_headers(self, conn, message): if len(message.headers): self.block_announced = True message.headers[-1].calc_sha256() self.last_blockhash_announced = message.headers[-1].sha256 # Test whether the last announcement we received had the # right header or the right inv # inv and headers should be lists of block hashes def check_last_announcement(self, headers=None, inv=None): expect_headers = headers if headers != None else [] expect_inv = inv if inv != None else [] test_function = lambda: self.block_announced wait_until(test_function, timeout=60, lock=mininode_lock) with mininode_lock: self.block_announced = False success = True compare_inv = [] if "inv" in self.last_message: compare_inv = [x.hash for x in self.last_message["inv"].inv] if compare_inv != expect_inv: success = False hash_headers = [] if "headers" in self.last_message: # treat headers as a list of block hashes hash_headers = [ x.sha256 for x in self.last_message["headers"].headers ] if hash_headers != expect_headers: success = False self.last_message.pop("inv", None) self.last_message.pop("headers", None) return success def wait_for_getdata(self, hash_list, timeout=60): if hash_list == []: return test_function = lambda: "getdata" in self.last_message and [x.hash for x in self.last_message["getdata"].inv] == hash_list wait_until(test_function, timeout=timeout, lock=mininode_lock) return def wait_for_block_announcement(self, block_hash, timeout=60): test_function = lambda: self.last_blockhash_announced == block_hash wait_until(test_function, timeout=timeout, lock=mininode_lock) return def send_header_for_blocks(self, new_blocks): headers_message = msg_headers() headers_message.headers = [ CBlockHeader(b) for b in new_blocks ] self.send_message(headers_message) def send_getblocks(self, locator): getblocks_message = msg_getblocks() getblocks_message.locator.vHave = locator self.send_message(getblocks_message) class SendHeadersTest(BitcoinTestFramework): def __init__(self): super().__init__() self.setup_clean_chain = True self.num_nodes = 2 # mine count blocks and return the new tip def mine_blocks(self, count): # Clear out last block announcement from each p2p listener [ x.clear_last_announcement() for x in self.p2p_connections ] self.nodes[0].generate(count) return int(self.nodes[0].getbestblockhash(), 16) # mine a reorg that invalidates length blocks (replacing them with # length+1 blocks). # Note: we clear the state of our p2p connections after the # to-be-reorged-out blocks are mined, so that we don't break later tests. # return the list of block hashes newly mined def mine_reorg(self, length): self.nodes[0].generate(length) # make sure all invalidated blocks are node0's sync_blocks(self.nodes, wait=0.1) for x in self.p2p_connections: x.wait_for_block_announcement(int(self.nodes[0].getbestblockhash(), 16)) x.clear_last_announcement() tip_height = self.nodes[1].getblockcount() hash_to_invalidate = self.nodes[1].getblockhash(tip_height-(length-1)) self.nodes[1].invalidateblock(hash_to_invalidate) all_hashes = self.nodes[1].generate(length+1) # Must be longer than the orig chain sync_blocks(self.nodes, wait=0.1) return [int(x, 16) for x in all_hashes] def run_test(self): # Setup the p2p connections and start up the network thread. inv_node = TestNode() test_node = TestNode() self.p2p_connections = [inv_node, test_node] connections = [] connections.append(NodeConn('127.0.0.1', p2p_port(0), self.nodes[0], inv_node)) # Set nServices to 0 for test_node, so no block download will occur outside of # direct fetching connections.append(NodeConn('127.0.0.1', p2p_port(0), self.nodes[0], test_node, services=0)) inv_node.add_connection(connections[0]) test_node.add_connection(connections[1]) NetworkThread().start() # Start up network handling in another thread # Test logic begins here inv_node.wait_for_verack() test_node.wait_for_verack() tip = int(self.nodes[0].getbestblockhash(), 16) # PART 1 # 1. Mine a block; expect inv announcements each time self.log.info("Part 1: headers don't start before sendheaders message...") for i in range(4): old_tip = tip tip = self.mine_blocks(1) assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(inv=[tip]), True) # Try a few different responses; none should affect next announcement if i == 0: # first request the block test_node.get_data([tip]) test_node.wait_for_block(tip) elif i == 1: # next try requesting header and block test_node.get_headers(locator=[old_tip], hashstop=tip) test_node.get_data([tip]) test_node.wait_for_block(tip) test_node.clear_last_announcement() # since we requested headers... elif i == 2: # this time announce own block via headers height = self.nodes[0].getblockcount() last_time = self.nodes[0].getblock(self.nodes[0].getbestblockhash())['time'] block_time = last_time + 1 new_block = create_block(tip, create_coinbase(height+1), block_time) new_block.solve() test_node.send_header_for_blocks([new_block]) test_node.wait_for_getdata([new_block.sha256]) test_node.send_message(msg_block(new_block)) test_node.sync_with_ping() # make sure this block is processed inv_node.clear_last_announcement() test_node.clear_last_announcement() self.log.info("Part 1: success!") self.log.info("Part 2: announce blocks with headers after sendheaders message...") # PART 2 # 2. Send a sendheaders message and test that headers announcements # commence and keep working. test_node.send_message(msg_sendheaders()) prev_tip = int(self.nodes[0].getbestblockhash(), 16) test_node.get_headers(locator=[prev_tip], hashstop=0) test_node.sync_with_ping() # Now that we've synced headers, headers announcements should work tip = self.mine_blocks(1) assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(headers=[tip]), True) height = self.nodes[0].getblockcount()+1 block_time += 10 # Advance far enough ahead for i in range(10): # Mine i blocks, and alternate announcing either via # inv (of tip) or via headers. After each, new blocks # mined by the node should successfully be announced # with block header, even though the blocks are never requested for j in range(2): blocks = [] for b in range(i+1): blocks.append(create_block(tip, create_coinbase(height), block_time)) blocks[-1].solve() tip = blocks[-1].sha256 block_time += 1 height += 1 if j == 0: # Announce via inv test_node.send_block_inv(tip) test_node.wait_for_getheaders() # Should have received a getheaders now test_node.send_header_for_blocks(blocks) # Test that duplicate inv's won't result in duplicate # getdata requests, or duplicate headers announcements [ inv_node.send_block_inv(x.sha256) for x in blocks ] test_node.wait_for_getdata([x.sha256 for x in blocks]) inv_node.sync_with_ping() else: # Announce via headers test_node.send_header_for_blocks(blocks) test_node.wait_for_getdata([x.sha256 for x in blocks]) # Test that duplicate headers won't result in duplicate # getdata requests (the check is further down) inv_node.send_header_for_blocks(blocks) inv_node.sync_with_ping() [ test_node.send_message(msg_block(x)) for x in blocks ] test_node.sync_with_ping() inv_node.sync_with_ping() # This block should not be announced to the inv node (since it also # broadcast it) assert "inv" not in inv_node.last_message assert "headers" not in inv_node.last_message tip = self.mine_blocks(1) assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(headers=[tip]), True) height += 1 block_time += 1 self.log.info("Part 2: success!") self.log.info("Part 3: headers announcements can stop after large reorg, and resume after headers/inv from peer...") # PART 3. Headers announcements can stop after large reorg, and resume after # getheaders or inv from peer. for j in range(2): # First try mining a reorg that can propagate with header announcement new_block_hashes = self.mine_reorg(length=7) tip = new_block_hashes[-1] assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(headers=new_block_hashes), True) block_time += 8 # Mine a too-large reorg, which should be announced with a single inv new_block_hashes = self.mine_reorg(length=8) tip = new_block_hashes[-1] assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(inv=[tip]), True) block_time += 9 fork_point = self.nodes[0].getblock("%02x" % new_block_hashes[0])["previousblockhash"] fork_point = int(fork_point, 16) # Use getblocks/getdata test_node.send_getblocks(locator = [fork_point]) assert_equal(test_node.check_last_announcement(inv=new_block_hashes), True) test_node.get_data(new_block_hashes) test_node.wait_for_block(new_block_hashes[-1]) for i in range(3): # Mine another block, still should get only an inv tip = self.mine_blocks(1) assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(inv=[tip]), True) if i == 0: # Just get the data -- shouldn't cause headers announcements to resume test_node.get_data([tip]) test_node.wait_for_block(tip) elif i == 1: # Send a getheaders message that shouldn't trigger headers announcements # to resume (best header sent will be too old) test_node.get_headers(locator=[fork_point], hashstop=new_block_hashes[1]) test_node.get_data([tip]) test_node.wait_for_block(tip) elif i == 2: test_node.get_data([tip]) test_node.wait_for_block(tip) # This time, try sending either a getheaders to trigger resumption # of headers announcements, or mine a new block and inv it, also # triggering resumption of headers announcements. if j == 0: test_node.get_headers(locator=[tip], hashstop=0) test_node.sync_with_ping() else: test_node.send_block_inv(tip) test_node.sync_with_ping() # New blocks should now be announced with header tip = self.mine_blocks(1) assert_equal(inv_node.check_last_announcement(inv=[tip]), True) assert_equal(test_node.check_last_announcement(headers=[tip]), True) self.log.info("Part 3: success!") self.log.info("Part 4: Testing direct fetch behavior...") tip = self.mine_blocks(1) height = self.nodes[0].getblockcount() + 1 last_time = self.nodes[0].getblock(self.nodes[0].getbestblockhash())['time'] block_time = last_time + 1 # Create 2 blocks. Send the blocks, then send the headers. blocks = [] for b in range(2): blocks.append(create_block(tip, create_coinbase(height), block_time)) blocks[-1].solve() tip = blocks[-1].sha256 block_time += 1 height += 1 inv_node.send_message(msg_block(blocks[-1])) inv_node.sync_with_ping() # Make sure blocks are processed test_node.last_message.pop("getdata", None) test_node.send_header_for_blocks(blocks) test_node.sync_with_ping() # should not have received any getdata messages with mininode_lock: assert "getdata" not in test_node.last_message # This time, direct fetch should work blocks = [] for b in range(3): blocks.append(create_block(tip, create_coinbase(height), block_time)) blocks[-1].solve() tip = blocks[-1].sha256 block_time += 1 height += 1 test_node.send_header_for_blocks(blocks) test_node.sync_with_ping() test_node.wait_for_getdata([x.sha256 for x in blocks], timeout=direct_fetch_response_time) [ test_node.send_message(msg_block(x)) for x in blocks ] test_node.sync_with_ping() # Now announce a header that forks the last two blocks tip = blocks[0].sha256 height -= 1 blocks = [] # Create extra blocks for later for b in range(20): blocks.append(create_block(tip, create_coinbase(height), block_time)) blocks[-1].solve() tip = blocks[-1].sha256 block_time += 1 height += 1 # Announcing one block on fork should not trigger direct fetch # (less work than tip) test_node.last_message.pop("getdata", None) test_node.send_header_for_blocks(blocks[0:1]) test_node.sync_with_ping() with mininode_lock: assert "getdata" not in test_node.last_message # Announcing one more block on fork should trigger direct fetch for # both blocks (same work as tip) test_node.send_header_for_blocks(blocks[1:2]) test_node.sync_with_ping() test_node.wait_for_getdata([x.sha256 for x in blocks[0:2]], timeout=direct_fetch_response_time) # Announcing 16 more headers should trigger direct fetch for 14 more # blocks test_node.send_header_for_blocks(blocks[2:18]) test_node.sync_with_ping() test_node.wait_for_getdata([x.sha256 for x in blocks[2:16]], timeout=direct_fetch_response_time) # Announcing 1 more header should not trigger any response test_node.last_message.pop("getdata", None) test_node.send_header_for_blocks(blocks[18:19]) test_node.sync_with_ping() with mininode_lock: assert "getdata" not in test_node.last_message self.log.info("Part 4: success!") # Now deliver all those blocks we announced. [ test_node.send_message(msg_block(x)) for x in blocks ] self.log.info("Part 5: Testing handling of unconnecting headers") # First we test that receipt of an unconnecting header doesn't prevent # chain sync. for i in range(10): test_node.last_message.pop("getdata", None) blocks = [] # Create two more blocks. for j in range(2): blocks.append(create_block(tip, create_coinbase(height), block_time)) blocks[-1].solve() tip = blocks[-1].sha256 block_time += 1 height += 1 # Send the header of the second block -> this won't connect. with mininode_lock: test_node.last_message.pop("getheaders", None) test_node.send_header_for_blocks([blocks[1]]) test_node.wait_for_getheaders() test_node.send_header_for_blocks(blocks) test_node.wait_for_getdata([x.sha256 for x in blocks]) [ test_node.send_message(msg_block(x)) for x in blocks ] test_node.sync_with_ping() assert_equal(int(self.nodes[0].getbestblockhash(), 16), blocks[1].sha256) blocks = [] # Now we test that if we repeatedly don't send connecting headers, we # don't go into an infinite loop trying to get them to connect. MAX_UNCONNECTING_HEADERS = 10 for j in range(MAX_UNCONNECTING_HEADERS+1): blocks.append(create_block(tip, create_coinbase(height), block_time)) blocks[-1].solve() tip = blocks[-1].sha256 block_time += 1 height += 1 for i in range(1, MAX_UNCONNECTING_HEADERS): # Send a header that doesn't connect, check that we get a getheaders. with mininode_lock: test_node.last_message.pop("getheaders", None) test_node.send_header_for_blocks([blocks[i]]) test_node.wait_for_getheaders() # Next header will connect, should re-set our count: test_node.send_header_for_blocks([blocks[0]]) # Remove the first two entries (blocks[1] would connect): blocks = blocks[2:] # Now try to see how many unconnecting headers we can send # before we get disconnected. Should be 5*MAX_UNCONNECTING_HEADERS for i in range(5*MAX_UNCONNECTING_HEADERS - 1): # Send a header that doesn't connect, check that we get a getheaders. with mininode_lock: test_node.last_message.pop("getheaders", None) test_node.send_header_for_blocks([blocks[i%len(blocks)]]) test_node.wait_for_getheaders() # Eventually this stops working. test_node.send_header_for_blocks([blocks[-1]]) # Should get disconnected test_node.wait_for_disconnect() self.log.info("Part 5: success!") # Finally, check that the inv node never received a getdata request, # throughout the test assert "getdata" not in inv_node.last_message if __name__ == '__main__': SendHeadersTest().main()