fad1de66a2 wallet: Remove unused boost::this_thread::interruption_point (MarcoFalke)
Pull request description:
`BerkeleyEnvironment::Open` is only called from the main thread (init) or an http rpc thread, neither of which can be interrupted, so remove the useless interruption point.
`BerkeleyEnvironment{}` is only used in tests, which run in a single process/thread, so remove the useless interruption point.
ACKs for top commit:
laanwj:
ACK fad1de66a2
fanquake:
ACK fad1de66a2
Tree-SHA512: dacd8398e966e4a6ce5cf7d3ed821c9c267eff40b14c0635085441647cdb72d1642807f89355419f1710f814c7963e35a10d102d0b985c7198261dfc736256f8
0b75a7f068 wallet: Reuse existing batch in CWallet::SetUsedDestinationState (João Barbosa)
01f45dd00e wallet: Avoid recursive lock in CWallet::SetUsedDestinationState (João Barbosa)
Pull request description:
This PR makes 2 distinct changes around `CWallet::SetUsedDestinationState`:
- 1st the recursive lock is removed and now it requires the lock to be held;
- 2nd change is to support, in the best case, just a wallet database flush when transaction is added to the wallet.
ACKs for top commit:
achow101:
ACK 0b75a7f068
MarcoFalke:
ACK 0b75a7f068
ryanofsky:
Code review ACK 0b75a7f068. Code changes looks fine but PR description should be updated to say what benefits of the change are. I might have missed something, but I didn't see a place where multiple batches were used previously and a single batch was used now. So the main benefit of this change appears to be removing a recursive lock? And maybe moving toward a consistent convention for passing batch instances?
Tree-SHA512: abcf23a5850d29990668db20d6f624cca3e89629cc9ed003e0d05cde1b58ab2ff365034f156684ad13e55764b54c6c0c2bc7d5f96b8af7dc5e45a3be955d6b15
36b68de5b2 Remove getBlockDepth method from Chain::interface (Antoine Riard)
b66c429c56 Remove locked_chain from GetDepthInMainChain and its callers (Antoine Riard)
0ff03871ad Use CWallet::m_last_block_processed_height in GetDepthInMainChain (Antoine Riard)
f77b1de16f Only return early from BlockUntilSyncedToCurrentChain if current tip is exact match (Antoine Riard)
769ff05e48 Refactor some importprunedfunds checks with guard clause (Antoine Riard)
5971d3848e Add block_height field in struct Confirmation (Antoine Riard)
9700fcb47f Replace CWalletTx::SetConf by Confirmation initialization list (Antoine Riard)
5aacc3eff1 Add m_last_block_processed_height field in CWallet (Antoine Riard)
10b4729e33 Pass block height in Chain::BlockConnected/Chain::BlockDisconnected (Antoine Riard)
Pull request description:
Work starter to remove Chain::Lock interface by adding m_last_block_processed_height in CWallet and m_block_height in CMerkleTx to avoid GetDepthInMainChain having to keep a lock . Once this one done, it should ease work to wipe out more cs_main locks from wallet code.
I think it's ready for a first round of review before to get further.
- `BlockUntilSyncedToCurrent` : restrain isPotentialTip to isTip because we want to be sure that wallet see BlockDisconnected callbacks if its height differs from the Chain one. It means during a reorg, an RPC could return before the BlockDisconnected callback had been triggered. This could cause a tx that had been included in the disconnected block to be displayed as confirmed, for example.
~~- `AbandonTransaction` : in case of conflicted tx (nIndex = -1), we set its m_block_height to the one of conflicting blocks, but if this height is superior to CWallet::m_last_block_processed_height, that means tx isn't conflicted anymore so we return 0 as tx is again unconfirmed~~ After #16624, we instead rely on Confirmation.
~~- `AddToWalletIfInvolvingMe`: in case of block disconnected, transactions are added to mempool again, so we need to replace old txn in `mapWallet` with a height set to zero so we remove check on block_hash.IsNull~~ Already done in #16624
ACKs for top commit:
jnewbery:
@jkczyz you've ACKed an intermediate commit (github annoyingly orders commits in date order, not commit order). Did you mean to ACK the final commit in this branch (36b68de5b2).
jkczyz:
> @jkczyz you've ACKed an intermediate commit (github annoyingly orders commits in date order, not commit order). Did you mean to ACK the final commit in this branch ([36b68de](36b68de5b2)).
meshcollider:
utACK 36b68de5b2
ryanofsky:
Code review ACK 36b68de5b2. Changes since last review: new jkczyz refactor importprunedfunds commit, changed BlockUntilSyncedToCurrentChainChanges commit title and description, changed Confirmation struct field order and line-wrapped comment
jnewbery:
utACK 36b68de5b2
promag:
Code review ACK 36b68de5b2.
Tree-SHA512: 08b89a0bcc39f67c82a6cb6aee195e6a11697770c788ba737b90986b4893f44e90d1ab9ef87239ea3766508b7e24ea882b7199df41173ab27a3d000328c14644
fa4c6fa9b1 doc: Add documentation for new test/lib (MarcoFalke)
faec28252c scripted-diff: test: Move setup_common to test library (MarcoFalke)
Pull request description:
Sorry for clickbait, this is only a move-only scripted-diff commit and one documentation commit.
Longer term, someone who knows something about build systems can make this an actual library. Motivation for this is that each module gets compiled for each target that includes it. For example, setup_common is compiled 27 times (for the fuzz suite) and another 3 times for the other tests (bench, unit test, gui)
ACKs for top commit:
practicalswift:
ACK fa4c6fa9b1 -- diff looks correct and Travis is happy
jonatack:
ACK fa4c6fa9b1 with the reserve that the commit messages (and PR description) contain the motivation for this change. Built, ran tests, light code review.
ryanofsky:
Code review ACK fa4c6fa9b1. I didn't realize `lib` was actually name of existing directory, not a new name. But in any case this looks good and nice to have one scripted diff instead of two.
Tree-SHA512: 2e176df90c60578276e4a6dc83ff57ff59d8e666ecf30c5ceacb8c326725da91baa4cac3dfa7a2e1605f58122a3e3e27e4938ff33e3a0ce7ea53afffebbf57a4
We don't remove yet Chain locks as we need to preserve lock
order with CWallet one until swapping at once to avoid
deadlock failures (spotted by --enable-debug)
is exact match
In the next commit, we start using BlockConnected/BlockDisconnected
callbacks to establish tx depth, rather than querying the chain
directly.
Currently, BlockUntilSyncedToCurrentChain will return early if
the best block processed by the wallet is a descendant of the node'tip.
That means that in the case of a re-org, it won't wait for the
BlockDisconnected callbacks that have been enqueued during the re-org
but have not yet been triggered in the wallet.
Change BlockUntilSyncedToCurrentChain to only return early if the
wallet's m_last_block_processed matches the tip exactly. This ensures
that there are no BlockDisconnected or BlockConnected callbacks
in-flight.
At wallet loading, we rely on chain state querying to retrieve
height of txn, to do so we ensure that lock order is respected
between cs_main and cs_wallet.
If wallet loaded is the wallet-tool one, all wallet txn will
show up with a height of zero. It doesn't matter as confirmation
height is not used by wallet-tool.
Reorder arguments and document Confirmation calls to avoid
ambiguity.
Fixes nits left from #16624
fa2c44c3cc test: Add ASSERT_DEBUG_LOG to unit test framework (MarcoFalke)
fa1936f57b logging: Add member for arbitrary print callbacks (MarcoFalke)
Pull request description:
Similar to `assert_debug_log` in the functional test framework
Top commit has no ACKs.
Tree-SHA512: aa9eaeca386b61d806867c04a33275f6eb4624fa5bf50f2928d16c83f5634bac96bcac46f9e8eda3b00b4251c5f12d7b01d6ffd84ba8e05c09eeec810cc31251
To do so we update CValidationInterface::BlockDisconnect to take a
CBlockIndex pointing to the block being disconnected.
This new parameter will be use in the following commit to establish
wallet height.
Suggested https://github.com/bitcoin/bitcoin/pull/17304#discussion_r341194391
by Gregory Sanders <gsanders87@gmail.com>
Reason for keeping the `return true` `return false` verbosity is that more code
will be added after the ReserveKeyFromKeyPool() call before returning.
4671fc3d9e Expand on wallet_balance.py comment from https://github.com/bitcoin/bitcoin/pull/16766\#issuecomment-527563982 (Jeremy Rubin)
91f3073f08 Update release notes to mention changes to IsTrusted and impact on wallet (Jeremy Rubin)
8f174ef112 Systematize style of IsTrusted single line if (Jeremy Rubin)
b49dcbedf7 update variable naming conventions for IsTrusted (Jeremy Rubin)
5ffe0d1449 Update comment in test/functional/wallet_balance.py (Jeremy Rubin)
a550c58267 Update wallet_balance.py test to reflect new behavior (Jeremy Rubin)
5dd7da4ccd Reuse trustedParents in looped calls to IsTrusted (Jeremy Rubin)
595f09d6de Cache tx Trust per-call to avoid DoS (Jeremy Rubin)
dce032ce29 Make IsTrusted scan parents recursively (Jeremy Rubin)
Pull request description:
This slightly modifies the behavior of IsTrusted to recursively check the parents of a transaction. Otherwise, it's possible that a parent is not IsTrusted but a child is. If a parent is not trusted, then a child should not be either.
This recursive scan can be a little expensive, so ~it might be beneficial to have a way of caching IsTrusted state, but this is a little complex because various conditions can change between calls to IsTrusted (e.g., re-org).~ I added a cache which works per call/across calls, but does not store the results semi-permanently. Which reduces DoS risk of this change. There is no risk of untrusted parents causing a resource exploitation, as we immediately return once that is detected.
This is a change that came up as a bug-fix esque change while working on OP_SECURETHEBAG. You can see the branch where this change is important here: https://github.com/bitcoin/bitcoin/compare/master...JeremyRubin:stb-with-rpc?expand=1. Essentially, without this change, we can be tricked into accepting an OP_SECURETHEBAG output because we don't properly check the parents. As this was a change which, on its own, was not dependent on OP_SECURETHEBAG, I broke it out as I felt the change stands on its own by fixing a long standing wallet bug.
The test wallet_balance.py has been corrected to meet the new behavior. The below comment, reproduced, explains what the issue is and the edge cases that can arise before this change.
# Before `test_balance()`, we have had two nodes with a balance of 50
# each and then we:
#
# 1) Sent 40 from node A to node B with fee 0.01
# 2) Sent 60 from node B to node A with fee 0.01
#
# Then we check the balances:
#
# 1) As is
# 2) With transaction 2 from above with 2x the fee
#
# Prior to #16766, in this situation, the node would immediately report
# a balance of 30 on node B as unconfirmed and trusted.
#
# After #16766, we show that balance as unconfirmed.
#
# The balance is indeed "trusted" and "confirmed" insofar as removing
# the mempool transactions would return at least that much money. But
# the algorithm after #16766 marks it as unconfirmed because the 'taint'
# tracking of transaction trust for summing balances doesn't consider
# which inputs belong to a user. In this case, the change output in
# question could be "destroyed" by replace the 1st transaction above.
#
# The post #16766 behavior is correct; we shouldn't be treating those
# funds as confirmed. If you want to rely on that specific UTXO existing
# which has given you that balance, you cannot, as a third party
# spending the other input would destroy that unconfirmed.
#
# For example, if the test transactions were:
#
# 1) Sent 40 from node A to node B with fee 0.01
# 2) Sent 10 from node B to node A with fee 0.01
#
# Then our node would report a confirmed balance of 40 + 50 - 10 = 80
# BTC, which is more than would be available if transaction 1 were
# replaced.
The release notes have been updated to note the new behavior.
ACKs for top commit:
ariard:
Code Review ACK 4671fc3, maybe extend DoS protection in a follow-up PR.
fjahr:
Code review ACK 4671fc3d9e
ryanofsky:
Code review ACK 4671fc3d9e. Changes since last review: 2 new commits adding suggested release note and python test comment, also a clean rebase with no changes to the earlier commits. The PR description is more comprehensive now, too. Looks good!
promag:
Code review ACK 4671fc3d9e.
Tree-SHA512: 6b183ff425304fef49724290053514cb2770f4a2350dcb83660ef24af5c54f7c4c2c345b0f62bba60eb2d2f70625ee61a7fab76a7f491bb5a84be5c4cc86b92f
436ad43643 Fix issue with conflicted mempool tx in listsinceblock (Adam Jonas)
Pull request description:
Closes#8752 by bringing back abandoned #10470.
This now checks that returned transactions are not conflicting with any transactions that are filtered out by the given blockhash and add a functional test to prevent this in the future.
For more context, #8757 was closed in favor of #10470.
ACKs for top commit:
instagibbs:
utACK 436ad43643
kallewoof:
utACK 436ad43643
jonatack:
I'm not qualifed to give an ACK here but 436ad43643 appears reasonable. Built/ran tests/verified that this test fails without the change in rpcwallet.cpp:
Tree-SHA512: 63d75cd3d3f19fc84dc38899b200c96179b82b24db263cd0116ee5b715265be647157855c2e35912d2fbc49c7b37db9375d6aab0ac672f0f09bece8431de5ea9
c98bd13e67 replace asserts in RPC code with CHECK_NONFATAL and add linter (Adam Jonas)
Pull request description:
- Replace instances of assert in /rpc files and rpcwallet with CHECK_NONFATAL(condition)
- Add a linter to prevent future usage of assert being used in RPC code
ref https://github.com/bitcoin/bitcoin/pull/17192
ACKs for top commit:
practicalswift:
ACK c98bd13e67 -- diff looks correct
Tree-SHA512: a16036b6bbcca73a5334665f66e17e1756377d582317568291da1d727fc9cf8c84bac9d9bd099534e1be315345336e5f7b66b93793135155f320dc5862a2d875
ScriptPubKeyMan is only using UnsetWalletFlagWithDB to unset the blank
wallet flag. Just make that it's own function and not expose the flag
writing directly.
This does not change behavior.
