Add preciousblock RPC #6996

sipa · Nov 12, 2015

This adds a new PRC 'preciousblock' to mark a block as precious. A precious block will be treated as if it was received earlier than any other competing block.

sipa · Nov 12, 2015

This was suggested in #6995.

kanoi · Nov 12, 2015

Awesome name :)

gmaxwell · Nov 12, 2015

Concept ACK.

This needs to document that overriding the first seen block behavior in slows network convergence and can increase the rate of long reorganizations.

Nit: What about being more verbose and return true? Considering the UX part of that, some users may interpret an empty line as "failed".

_{Further reading: http://ux.stackexchange.com/a/12819}

It's very common to return Null in case there is no real result in our API.

Some places also use return true. But yes, null is more common.

new issue: #7043

TheBlueMatt · Nov 12, 2015

I would prefer if it would only prefer the given block over other blocks that arrived within, eg, 10 seconds of the other block in question.

On November 12, 2015 3:59:35 AM PST, Pieter Wuille notifications@github.com wrote:

This adds a new PRC 'preciousblock' to mark a block as precious. A
precious block will be treated as if it was received earlier than any
other competing block.
You can view, comment on, or merge this pull request online at:

#6996

-- Commit Summary --

Add preciousblock RPC

-- File Changes --

M qa/pull-tester/rpc-tests.py (1)
A qa/rpc-tests/preciousblock.py (66)
M src/chain.h (2)
M src/main.cpp (15)
M src/main.h (3)
M src/rpcblockchain.cpp (38)

-- Patch Links --

https://github.com/bitcoin/bitcoin/pull/6996.patch
https://github.com/bitcoin/bitcoin/pull/6996.diff

Reply to this email directly or view it on GitHub:
#6996

sipa · Nov 12, 2015

@TheBlueMatt That would be a much larger change, as we don't track receive time.

TheBlueMatt · Nov 12, 2015

Indeed, not sure it's worth it, but would prefer if we could :).

On November 12, 2015 12:33:08 PM PST, Pieter Wuille notifications@github.com wrote:

@TheBlueMatt That would be a much larger change, as we don't track
receive time.

Reply to this email directly or view it on GitHub:
#6996 (comment)

MarcoFalke · Nov 12, 2015

utACK

gmaxwell · Nov 13, 2015

I'm sympathetic with matt's ask; but I don't think it's worth it-- it'll get bypassed (which then puts parties back into a mode of modifying the consensus code), and it would imply that 10 seconds was "safe"-- I have no idea if it is... the inherent preference for your own blocks by virtue of hearing the first already hurts convergence; so I think it would be fair to say that no amount additional is "safe"... but having a well constructed option is much safer than people trying to achieve this with invalidateblock.

kanoi · Nov 13, 2015

Well any numerical limitation would of course only require a few bytes of code change to circumvent so would seem to be somewhat a waste of time considering the extra work involved: as @sipa suggests

sipa · Nov 28, 2015

@gmaxwell What do you mean by 'saturate' ?

sipa · Nov 28, 2015

Ok, changed the approach a bit. It now only keeps a counter per tip, and resets it when more work was added.

gmaxwell · Nov 28, 2015

By saturate I mean what you did, sat_add(int_min, -1) = int_min

sipa · Nov 30, 2015

Rebased.

jgarzik · Nov 30, 2015

ut ACK

sipa · Nov 30, 2015

There is a bug in the unit tests that can lead to a race condition.

sipa · Dec 1, 2015

Fixed.

sipa · Dec 11, 2015

Rebased.

FYI, i don't think this call to sync_blocks is doing anything helpful -- it is just checking that getblockcount is the same between the two nodes, which it will be immediately, even before any headers or blocks are sent between the two nodes.

Perhaps this should sync on getchaintips() returning the expected output?

There's a race condition here; it's possible for nodes[1] to have not yet received hashC at the time this preciousblock call happens, which will cause the sync_chain in the next line to hang indefinitely.

(I had this happen to me on one of my runs of this test.)

laanwj · Feb 1, 2016

Needs rebase

Extra space at the end of the line.

luke-jr · Feb 26, 2016

This can break CheckBlockIndex:

if (pindex->nChainTx == 0) assert(pindex->nSequenceId == 0);  // nSequenceId can't be set for blocks that aren't linked

This assumption is no longer true if we use preciousblock on a not-yet-downloaded block (and this can occur in the RPC test).

gmaxwell · Apr 19, 2016

Needs rebase.

Needs to be changed to python3 after rebase.

Also some other adjustments are needed in this file. (I am happy to do those)

btcdrak · Jun 14, 2016

needs rebase

luke-jr · Aug 9, 2016

Pushed a rebase to my repo

laanwj · Aug 19, 2016

Are you still planning to do this?

sipa · Aug 26, 2016

@luke-jr Your rebase does not fix the comment you made here: #6996 (comment) ?

sipa · Aug 26, 2016

@luke-jr Switched to your commit e7a3457, and changed the test to python3 and updated the assert. The tests don't run, though. I'll inventigate later.

I think you need to use __init__() here, instead.

luke-jr · Aug 26, 2016

@sipa It is fixed in eafeb97, and the tests updated in 5e6af82 (both of which are part of my "preciousblock" branch)

sipa · Aug 26, 2016

@luke-jr Ah, a link to that would have been useful :)

I've squashed the fixes and split into 2 commits.

laanwj · Oct 18, 2016

utACK 5805ac8

btcdrak · Oct 19, 2016

utACK 5805ac8

rebroad · Oct 19, 2016

that was a surprise merge! Are there any plans that anyone knows of to use this? I would love to see some use-cases for documentation purposes.

laanwj added RPC/REST/ZMQ Validation labels Nov 12, 2015

sipa referenced this pull request Nov 12, 2015
Open
the RPC should have commands to specify the side of a fork considered 'current' #6995

luke-jr added a commit to luke-jr/bitcoin that referenced this pull request Feb 13, 2016

Add preciousblock RPC
Github-Pull: #6996 Rebased-From: 6fda880

1683eb9

laanwj added the Feature label Jun 16, 2016

luke-jr added a commit to bitcoinknots/bitcoin that referenced this pull request Jun 28, 2016

Merge #6996 preciousblock-0.12.x

bb7b476

luke-jr added a commit to bitcoinknots/bitcoin that referenced this pull request Aug 6, 2016

Add preciousblock RPC
Github-Pull: #6996 Rebased-From: 6fda880

e7a3457

sipa added some commits Aug 26, 2016

Add preciousblock RPC
Includes a bugfix by Luke-Jr.

5127c4f

Add preciousblock tests
Rebased, improved and extended by Luke-Jr.

5805ac8

laanwj merged commit 5805ac8 into bitcoin:master Oct 18, 2016
1 check passed

1 check passed

continuous-integration/travis-ci/pr The Travis CI build passed
Details

This was referenced Oct 20, 2016

Closed

TODO for release notes 0.14.0 #8455

Open

Prioritize submitblock blocks to prevent attacks on miners using slow-to-verify blocks #5538

codablock added a commit to codablock/dash that referenced this pull request Sep 19, 2017

Merge #6996: Add preciousblock RPC
5805ac8 Add preciousblock tests (Pieter Wuille) 5127c4f Add preciousblock RPC (Pieter Wuille)

23bf8a6

+    +    {
+    +        LOCK(cs_main);
+    +        if (mapBlockIndex.count(hash) == 0)
+    +            throw JSONRPCError(RPC_INVALID_ADDRESS_OR_KEY, "Block not found");
+    +
+    +        pblockindex = mapBlockIndex[hash];
+    +    }
+    +
+    +    CValidationState state;
+    +    PreciousBlock(state, Params().GetConsensus(), pblockindex);
+    +
+    +    if (!state.IsValid()) {
+    +        throw JSONRPCError(RPC_DATABASE_ERROR, state.GetRejectReason());
+    +    }
+    +
+    +    return NullUniValue;

+    +        self.nodes.append(start_node(1, self.options.tmpdir, ["-debug"]))
+    +        self.nodes.append(start_node(2, self.options.tmpdir, ["-debug"]))
+    +
+    +    def run_test(self):
+    +        print "Mine blocks A-B-C on Node 0"
+    +        self.nodes[0].generate(3)
+    +        assert(self.nodes[0].getblockcount() == 3)
+    +        hashC = self.nodes[0].getbestblockhash()
+    +        print "Mine competing blocks E-F-G on Node 1"
+    +        self.nodes[1].generate(3)
+    +        assert(self.nodes[1].getblockcount() == 3)
+    +        hashG = self.nodes[1].getbestblockhash()
+    +        assert(hashC != hashG)
+    +        print "Connect nodes and check no reorg occurs"
+    +        connect_nodes_bi(self.nodes,0,1)
+    +        sync_blocks(self.nodes[0:2])

+    +        assert(self.nodes[1].getblockcount() == 3)
+    +        hashG = self.nodes[1].getbestblockhash()
+    +        assert(hashC != hashG)
+    +        print "Connect nodes and check no reorg occurs"
+    +        connect_nodes_bi(self.nodes,0,1)
+    +        sync_blocks(self.nodes[0:2])
+    +        assert(self.nodes[0].getbestblockhash() == hashC)
+    +        assert(self.nodes[1].getbestblockhash() == hashG)
+    +        print "Make Node0 prefer block G"
+    +        self.nodes[0].preciousblock(hashG)
+    +        assert(self.nodes[0].getbestblockhash() == hashG)
+    +        print "Make Node0 prefer block C again"
+    +        self.nodes[0].preciousblock(hashC)
+    +        assert(self.nodes[0].getbestblockhash() == hashC)
+    +        print "Make Node1 prefer block C"
+    +        self.nodes[1].preciousblock(hashC)

+    +
+    +#
+    +# Test PreciousBlock code
+    +#
+    +
+    +from test_framework.test_framework import BitcoinTestFramework
+    +from test_framework.util import *
+    +
+    +class PreciousTest(BitcoinTestFramework):
+    +    def setup_chain(self):
+    +        print("Initializing test directory "+self.options.tmpdir)
+    +        initialize_chain_clean(self.options.tmpdir, 3)
+    +
+    +    def setup_network(self):
+    +        self.nodes = []
+    +        self.is_network_split = False

@@ -0,0 +1,87 @@
+    +#!/usr/bin/env python3
+    +# Copyright (c) 2015 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 PreciousBlock code
+    +#
+    +
+    +from test_framework.test_framework import BitcoinTestFramework
+    +from test_framework.util import *
+    +
+    +class PreciousTest(BitcoinTestFramework):
+    +    def setup_chain(self):

bitcoin/bitcoin

Join GitHub today

Add preciousblock RPC #6996

Conversation

sipa commented Nov 12, 2015

laanwj added RPC/REST/ZMQ Validation labels Nov 12, 2015

sipa commented Nov 12, 2015

sipa referenced this pull request Nov 12, 2015

kanoi commented Nov 12, 2015

gmaxwell commented Nov 12, 2015

Show 4 comments Hide comments MarcoFalke and 1 other commented on an outdated diff Nov 12, 2015

TheBlueMatt commented Nov 12, 2015

sipa commented Nov 12, 2015

TheBlueMatt commented Nov 12, 2015

MarcoFalke commented Nov 12, 2015

gmaxwell commented Nov 13, 2015

kanoi commented Nov 13, 2015

sipa commented Nov 28, 2015

sipa commented Nov 28, 2015

gmaxwell commented Nov 28, 2015

sipa commented Nov 30, 2015

jgarzik commented Nov 30, 2015

sipa commented Nov 30, 2015

sipa commented Dec 1, 2015

sipa commented Dec 11, 2015

Show 1 comment Hide comment sdaftuar commented on an outdated diff Dec 14, 2015

Show 1 comment Hide comment sdaftuar commented on an outdated diff Dec 14, 2015

laanwj commented Feb 1, 2016

Show 1 comment Hide comment luke-jr commented on an outdated diff Feb 12, 2016

luke-jr added a commit to luke-jr/bitcoin that referenced this pull request Feb 13, 2016

luke-jr commented Feb 26, 2016

gmaxwell commented Apr 19, 2016

Show 1 comment Hide comment MarcoFalke commented on an outdated diff Jun 3, 2016

MarcoFalke Jun 3, 2016 • edited Edited 1 time MarcoFalke Jun 3, 2016

btcdrak commented Jun 14, 2016

laanwj added the Feature label Jun 16, 2016

luke-jr added a commit to bitcoinknots/bitcoin that referenced this pull request Jun 28, 2016

luke-jr added a commit to bitcoinknots/bitcoin that referenced this pull request Aug 6, 2016

luke-jr commented Aug 9, 2016

laanwj commented Aug 19, 2016

sipa commented Aug 26, 2016

sipa commented Aug 26, 2016

Show 1 comment Hide comment MarcoFalke commented on an outdated diff Aug 26, 2016

luke-jr commented Aug 26, 2016

sipa added some commits Aug 26, 2016

sipa commented Aug 26, 2016

laanwj commented Oct 18, 2016

Hide details View details laanwj merged commit 5805ac8 into bitcoin:master Oct 18, 2016 1 check passed

1 check passed

laanwj added a commit that referenced this pull request Oct 18, 2016

btcdrak commented Oct 19, 2016

rebroad commented Oct 19, 2016

This was referenced Oct 20, 2016

codablock added a commit to codablock/dash that referenced this pull request Sep 19, 2017

MarcoFalke and 1 other commented on an outdated diff Nov 12, 2015

sdaftuar commented on an outdated diff Dec 14, 2015

sdaftuar commented on an outdated diff Dec 14, 2015

luke-jr commented on an outdated diff Feb 12, 2016

MarcoFalke commented on an outdated diff Jun 3, 2016

MarcoFalke Jun 3, 2016 •

Edited 1 time

MarcoFalke
Jun 3, 2016

MarcoFalke commented on an outdated diff Aug 26, 2016

laanwj merged commit `5805ac8` into bitcoin:master Oct 18, 2016
1 check passed