p2p: Disable BIP 61 by default #14054

pull MarcoFalke wants to merge 2 commits into bitcoin:master from MarcoFalke:Mf1808-noBIP61 changing 2 files +11 −1
  1. MarcoFalke commented at 8:23 pm on August 24, 2018: member

    The live p2p network should not be used for debugging or as development aid when implementing the p2p protocol. Instead, applications should be tested locally (e.g. by inspecting the debug log of a validating node on the local network)

    Using the p2p network for this purpose seems wasteful and even dangerous, as peers can not be trusted to send the correct reject messages or a reject message at all.

  2. MarcoFalke added the label P2P on Aug 24, 2018
  3. MarcoFalke force-pushed on Aug 24, 2018
  4. MarcoFalke added the label Needs release notes on Aug 24, 2018
  5. jnewbery commented at 8:41 pm on August 24, 2018: member

    Concept ACK!

    The changes to the test framework remove all testing of the p2p REJECT message. Shouldn’t we maintain at least some minor testing of REJECT, unless(/until) REJECT is removed entirely?

  6. MarcoFalke force-pushed on Aug 24, 2018
  7. MarcoFalke force-pushed on Aug 24, 2018
  8. gmaxwell commented at 10:19 pm on August 24, 2018: contributor
    Concept ACK.
  9. MarcoFalke force-pushed on Aug 24, 2018
  10. MarcoFalke commented at 10:37 pm on August 24, 2018: member
    @jnewbery Added back one test case for -enablebip61
  11. laanwj commented at 9:16 am on August 27, 2018: member

    Concept ACK

    I’m not convinced about removing REJECT completely, I think it is useful for testing and development (and have used it for that purpose a few times, for example during development of bitcoin-submittx).

    However I fully agree on the reasoning behind disabling it by default.

  12. in test/functional/test_framework/mininode.py:488 in fa17b5ac94 outdated
    484@@ -492,66 +485,59 @@ def send_blocks_and_test(self, blocks, rpc, success=True, request_block=True, re
    485            ensure that any getdata messages are responded to
    486          - if success is True: assert that the node's tip advances to the most recent block
    487          - if success is False: assert that the node's tip doesn't advance
    488-         - if reject_code and reject_reason are set: assert that the correct reject message is received"""
    489+         - if reject_code and reject_reason are set: assert that the correct reject message is logged"""
    


    jnewbery commented at 2:53 pm on August 27, 2018:
    s/reject_code and reject_reason are/reject_reason is
  13. in test/functional/test_framework/mininode.py:519 in fa17b5ac94 outdated
    531          - add all txs to our tx_store
    532          - send tx messages for all txs
    533          - if success is True/False: assert that the txs are/are not accepted to the mempool
    534          - if expect_disconnect is True: Skip the sync with ping
    535-         - if reject_code and reject_reason are set: assert that the correct reject message is received."""
    536+         - if reject_code and reject_reason are set: assert that the correct reject message is logged."""
    


    jnewbery commented at 2:56 pm on August 27, 2018:
    s/reject_code and reject_reason are/reject_reason is
  14. in test/functional/p2p_invalid_tx.py:146 in fa17b5ac94 outdated
    142@@ -143,9 +143,6 @@ def run_test(self):
    143                 "disconnecting peer=0",
    144         ]):
    145             node.p2p.send_txs_and_test([tx1], node, success=False, expect_disconnect=True)
    146-        # send_txs_and_test will have waited for disconnect, so we can safely check that no reject has been received
    


    jnewbery commented at 3:13 pm on August 27, 2018:

    This whole sub-test can be removed (everything from # restart node with sending BIP61 messages disabled, check that it disconnects without sending the reject message onward)

    You could also update L71 higher up to test the reject reason:

    0+++ b/test/functional/p2p_invalid_tx.py
    1@@ -68,7 +68,8 @@ class InvalidTxRequestTest(BitcoinTestFramework):
    2         # and we get disconnected immediately
    3         self.log.info('Test a transaction that is rejected')
    4         tx1 = create_tx_with_script(block1.vtx[0], 0, script_sig=b'\x64' * 35, amount=50 * COIN - 12000)
    5-        node.p2p.send_txs_and_test([tx1], node, success=False, expect_disconnect=True)
    6+        node.p2p.send_txs_and_test([tx1], node, success=False, expect_disconnect=True, 
    7+                                   reject_reason="{} from peer=0 was not accepted: mandatory-script-verify-flag-failed (Invalid OP_IF construction) (code 16)".format(tx1.hash))
    
  15. in test/functional/feature_dersig.py:121 in fa17b5ac94 outdated
    128-            if self.nodes[0].p2p.last_message["reject"].code == REJECT_INVALID:
    129-                # Generic rejection when a block is invalid
    130-                assert_equal(self.nodes[0].p2p.last_message["reject"].reason, b'block-validation-failed')
    131-            else:
    132-                assert b'Non-canonical DER signature' in self.nodes[0].p2p.last_message["reject"].reason
    133+            assert b'Non-canonical DER signature' in self.nodes[0].p2p.last_message["reject"].reason
    


    jnewbery commented at 3:14 pm on August 27, 2018:
    I have a mild preference to test reject messages in their own test case (p2p_reject.py), but not strongly enough to block this PR.
  16. in test/functional/feature_csv_activation.py:168 in fa17b5ac94 outdated
    164@@ -165,11 +165,11 @@ def create_test_block(self, txs, version=536870912):
    165         block.solve()
    166         return block
    167 
    168-    def sync_blocks(self, blocks, success=True, reject_code=None, reject_reason=None, request_block=True):
    169+    def sync_blocks(self, blocks, success=True, *, reject_reason=None, request_block=True):
    


    jnewbery commented at 3:18 pm on August 27, 2018:

    None of the calls to this function actually use the optional parameters, so this can be simplified to:

     0+++ b/test/functional/feature_csv_activation.py
     1@@ -165,11 +165,11 @@ class BIP68_112_113Test(BitcoinTestFramework):
     2         block.solve()
     3         return block
     4 
     5-    def sync_blocks(self, blocks, success=True, *, reject_reason=None, request_block=True):
     6+    def sync_blocks(self, blocks, success=True):
     7         """Sends blocks to test node. Syncs and verifies that tip has advanced to most recent block.
     8 
     9         Call with success = False if the tip shouldn't advance to the most recent block."""
    10-        self.nodes[0].p2p.send_blocks_and_test(blocks, self.nodes[0], success=success, reject_reason=reject_reason, request_block=request_block, expect_disconnect=False, timeout=60)
    11+        self.nodes[0].p2p.send_blocks_and_test(blocks, self.nodes[0], success=success)
    
  17. in test/functional/feature_cltv.py:62 in fa17b5ac94 outdated
    54@@ -51,10 +55,11 @@ def cltv_validate(node, tx, height):
    55                                   list(CScript(new_tx.vin[0].scriptSig)))
    56     return new_tx
    57 
    58+
    59 class BIP65Test(BitcoinTestFramework):
    60     def set_test_params(self):
    61         self.num_nodes = 1
    62-        self.extra_args = [['-whitelist=127.0.0.1']]
    63+        self.extra_args = [['-whitelist=127.0.0.1', '-par=1']]
    


    jnewbery commented at 3:25 pm on August 27, 2018:
    Please add a comment to explain why par=1 is needed.
  18. in test/functional/feature_block.py:172 in fa17b5ac94 outdated
    168@@ -169,7 +169,7 @@ def run_test(self):
    169         self.log.info("Reject a block where the miner creates too much coinbase reward")
    170         self.move_tip(6)
    171         b9 = self.next_block(9, spend=out[4], additional_coinbase_value=1)
    172-        self.sync_blocks([b9], False, 16, b'bad-cb-amount', reconnect=True)
    173+        self.sync_blocks([b9], False, 'coinbase pays too much (actual=10000000000 vs limit=9999999999)', reconnect=True)
    


    jnewbery commented at 3:30 pm on August 27, 2018:
    Would be nice if all calls to sync_blocks used named argument for reject_reason, but that change should be in a future PR.
  19. jnewbery commented at 3:32 pm on August 27, 2018: member
    Looks good. A few comments inline.
  20. MarcoFalke force-pushed on Aug 27, 2018
  21. DrahtBot commented at 5:32 pm on August 29, 2018: member
  22. MarcoFalke force-pushed on Aug 30, 2018
  23. luke-jr commented at 12:48 pm on August 30, 2018: member
    Reject may be useful (necessary) for tracking the best-block of peers and making sure our “essential” peers are using the same consensus rules as we are. Until we have reliable peer best-block tracking implemented, I suggest we leave reject alone.
  24. MarcoFalke force-pushed on Aug 31, 2018
  25. MarcoFalke force-pushed on Aug 31, 2018
  26. jnewbery commented at 8:51 pm on September 6, 2018: member

    Reject may be useful (necessary) for tracking the best-block of peers and making sure our “essential” peers are using the same consensus rules as we are.

    bitcoind does not track peers’ best blocks or consensus rules using REJECT messages. In fact, all it does with REJECT messages is log (if NET logging is enabled) and drop.

    I don’t think it makes sense to continue sending REJECT messages on the basis that they ‘may’ be used for something in future.

    This breaks on master in feature_block.py. You need this change:

    0@@ -824,7 +824,7 @@ class FullBlockTest(BitcoinTestFramework):
    1         tx.vin.append(CTxIn(COutPoint(b64a.vtx[1].sha256, 0)))
    2         b64a = self.update_block("64a", [tx])
    3         assert_equal(len(b64a.serialize()), MAX_BLOCK_BASE_SIZE + 8)
    4-        self.sync_blocks([b64a], success=False, reject_code=1, reject_reason=b'error parsing message')
    5+        self.sync_blocks([b64a], success=False, reject_reason='non-canonical ReadCompactSize(): iostream error')
    
  27. MarcoFalke commented at 9:40 pm on September 6, 2018: member

    @jnewbery I will rebase this on

    • #14119 (qa: Read reject reasons from debug log, not p2p messages by MarcoFalke)

    to get rid of all test changes and leave this a one-line change. Please help review that first.

  28. jnewbery commented at 11:42 am on September 7, 2018: member

    Please help review that first.

    Yes, fine (although a note to say that this PR isn’t seeking review would have been helpful!)

  29. p2p: Disable BIP 61 by default fa14b54a87
  30. MarcoFalke force-pushed on Sep 7, 2018
  31. MarcoFalke force-pushed on Sep 8, 2018
  32. doc: release notes for -enablebip61 default change faea5bfc5a
  33. MarcoFalke removed the label Needs release notes on Sep 9, 2018
  34. ken2812221 commented at 2:15 am on September 10, 2018: contributor
    utACK faea5bf
  35. jnewbery commented at 2:26 pm on September 10, 2018: member
    tested ACK faea5bfc5a975874acf763082852ed532ed81a95
  36. laanwj merged this on Sep 10, 2018
  37. laanwj closed this on Sep 10, 2018

  38. laanwj referenced this in commit 321075609d on Sep 10, 2018
  39. MarcoFalke deleted the branch on Sep 10, 2018
  40. MarcoFalke referenced this in commit d3a0382007 on Mar 14, 2019
  41. laanwj referenced this in commit 165ea14efe on Mar 16, 2019
  42. deadalnix referenced this in commit e269cb2f70 on Jun 9, 2020
  43. PastaPastaPasta referenced this in commit 08a9be48be on Jun 27, 2021
  44. PastaPastaPasta referenced this in commit 18ceaa64a5 on Jun 28, 2021
  45. PastaPastaPasta referenced this in commit 99326c32f5 on Jun 29, 2021
  46. Munkybooty referenced this in commit f73e76cf78 on Jul 1, 2021
  47. Munkybooty referenced this in commit 846f66fd75 on Sep 21, 2021
  48. Munkybooty referenced this in commit 943943399f on Sep 21, 2021
  49. Munkybooty referenced this in commit f983872ff2 on Sep 27, 2021
  50. MarcoFalke locked this on Dec 16, 2021

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin/bitcoin. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2024-11-17 09:12 UTC

This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me