3fcb545ab2 bench: benchmark transaction creation process (furszy)
a8a75346d7 wallet: SelectCoins, return early if target is covered by preset-inputs (furszy)
f41712a734 wallet: simplify preset inputs selection target check (furszy)
5baedc3351 wallet: remove fetch pre-selected-inputs responsibility from SelectCoins (furszy)
295852f619 wallet: encapsulate pre-selected-inputs lookup into its own function (furszy)
37e7887cb4 wallet: skip manually selected coins from 'AvailableCoins' result (furszy)
94c0766b0c wallet: skip available coins fetch if "other inputs" are disallowed (furszy)
Pull request description:
#### # Context (Current Flow on Master)
In the transaction creation process, in order to select which coins the new transaction will spend,
we first obtain all the available coins known by the wallet, which means walking-through the
wallet txes map, gathering the ones that fulfill certain spendability requirements in a vector.
This coins vector is then provided to the Coin Selection process, which first checks if the user
has manually selected any input (which could be internal, aka known by the wallet, or external),
and if it does, it fetches them by searching each of them inside the wallet and/or inside the
Coin Control external tx data.
Then, after finding the pre-selected-inputs and gathering them in a vector, the Coin Selection
process walks-through the entire available coins vector once more just to erase coins that are
in both vectors. So the Coin Selection process doesn’t pick them twice (duplicate inputs inside
the same transaction).
#### # Process Workflow Changes
Now, a new method, `FetchCoins` will be responsible for:
1) Lookup the user pre-selected-inputs (which can be internal or external).
2) And, fetch the available coins in the wallet (excluding the already fetched ones).
Which will occur prior to the Coin Selection process. Which allows us to never include the
pre-selected-inputs inside the available coins vector in the first place, as well as doing other
nice improvements (written below).
So, Coin Selection can perform its main responsibility without mixing it with having to fetch
internal/external coins nor any slow and unneeded duplicate coins verification.
#### # Summarizing the Improvements:
1) If any pre-selected-input lookup fail, the process will return the error right away.
(before, the wallet was fetching all the wallet available coins, walking through the
entire txes map, and then failing for an invalid pre-selected-input inside SelectCoins)
2) The pre-selected-inputs lookup failure causes are properly described on the return error.
(before, we were returning an "Insufficient Funds" error for everything, even if the failure
was due a not solvable external input)
3) **Faster Coin Selection**: no longer need to "remove the pre-set inputs from the available coins
vector so that Coin Selection doesn't pick them" (which meant to loop-over the entire
available coins vector at Coin Selection time, erasing duplicate coins that were pre-selected).
Now, the available coins vector, which is built after the pre-selected-inputs fetching,
doesn’t include the already selected inputs in the first place.
4) **Faster transaction creation** for transactions that only use manually selected inputs.
We now will return early, as soon as we finish fetching the pre-selected-inputs and
not perform the resources expensive calculation of walking-through the entire wallet
txes map to obtain the available coins (coins that we will not use).
---------------------------
Added a new bench (f6d0bb2) measuring the transaction creation process, for a wallet with ~250k UTXO, only using the pre-selected-inputs inside coin control. Setting `m_allow_other_inputs=false` to disallow the wallet to include coins automatically.
#### Result on this PR (tip f6d0bb2d):
| ns/op | op/s | err% | total | benchmark
|--------------------:|--------------------:|--------:|----------:|:----------
| 1,048,675.00 | 953.58 | 0.3% | 0.06 | `WalletCreateTransaction`
vs
#### Result on master (tip 4a4289e2):
| ns/op | op/s | err% | total | benchmark
|--------------------:|--------------------:|--------:|----------:|:----------
| 96,373,458.20 | 10.38 | 0.2% | 5.30 | `WalletCreateTransaction`
The benchmark took to run in master: **96.37 milliseconds**, while in this PR: **1 millisecond** 🚀 .
ACKs for top commit:
S3RK:
Code Review ACK 3fcb545ab2
achow101:
ACK 3fcb545ab2
aureleoules:
reACK 3fcb545ab2
Tree-SHA512: 42f833e92f40c348007ca565a4c98039e6f1ff25d8322bc2b27115824744779baf0b0a38452e4e2cdcba45076473f1028079bbd0f670020481ec5d3db42e4731
Goal 1:
Benchmark the transaction creation process for pre-selected-inputs only.
Setting `m_allow_other_inputs=false` to disallow the wallet to include coins automatically.
Goal 2:
Benchmark the transaction creation process for pre-selected-inputs and coin selection.
-----------------------
Benchmark Setup:
1) Generates a 5k blockchain, loading the wallet with 5k transactions with two outputs each.
2) Fetch 4 random UTXO from the wallet's available coins and pre-select them as inputs inside CoinControl.
Benchmark (Goal 1):
Call `CreateTransaction` providing the coin control, who has set `m_allow_other_inputs=false` and
the manually selected coins.
Benchmark (Goal 2):
Call `CreateTransaction` providing the coin control, who has set `m_allow_other_inputs=true` and
the manually selected coins.
aaaa7bd0ba iwyu: Add missing includes (MacroFake)
fa9ebec096 Remove g_parallel_script_checks (MacroFake)
fa7c834b9f Move ::fCheckBlockIndex into ChainstateManager (MacroFake)
fa43188d86 Move ::fCheckpointsEnabled into ChainstateManager (MacroFake)
cccca83099 Move ::nMinimumChainWork into ChainstateManager (MacroFake)
fa29d0b57c Move ::hashAssumeValid into ChainstateManager (MacroFake)
faf44876db Move ::nMaxTipAge into ChainstateManager (MacroFake)
Pull request description:
It seems preferable to assign globals to a class (in this case `ChainstateManager`), than to leave them dangling. This should clarify scope for code-readers, as well as clarifying unit test behaviour.
ACKs for top commit:
dergoegge:
Code review ACK aaaa7bd0ba
ryanofsky:
Code review ACK aaaa7bd0ba. No changes since last review, other than rebase
aureleoules:
reACK aaaa7bd0ba
Tree-SHA512: 83ec3ba0fb4f1dad95810d4bd4e578454e0718dc1bdd3a794cc4e48aa819b6f5dad4ac4edab3719bdfd5f89cbe23c2740a50fd56c1ff81c99e521c5f6d4e898d
SkipTo() reads data from the file into the CBufferedFile object
(memory), but, unlike this object's read() method, SkipTo() doesn't
transfer data into a caller's memory buffer. This is useful because
after skipping forward in the stream in this way, the user can, if
needed, rewind the stream (SetPos()) and access the object's memory
buffer including ranges that were skipped over (without needing to
read from the disk file).
This changes the flag for the bitcoin-chainstate executable. Previously
it was false, now it is the chain's default value (still false for the
main chain).
If we're connecting to the peer which might support
transaction reconciliation, we announce we want to reconcile
with them.
We store the reconciliation salt so that when the peer
responds with their salt, we are able to compute the
full reconciliation salt.
This behavior is enabled with a CLI flag.
626b7c8493 fuzz: add scanblocks as safe for fuzzing (James O'Beirne)
94fe5453c7 test: rpc: add scanblocks functional test (Jonas Schnelli)
6ef2566b68 rpc: add scanblocks - scan for relevant blocks with descriptors (Jonas Schnelli)
a4258f6e81 rpc: move-only: consolidate blockchain scan args (James O'Beirne)
Pull request description:
Revives #20664. All feedback from the previous PR has either been responded to inline or incorporated here.
---
Major changes from Jonas' PR:
- consolidated arguments for scantxoutset/scanblocks
- substantial cleanup of the functional test
Here's the range-diff (`git range-diff master jonasschnelli/2020/12/filterblocks_rpc jamesob/2021-11-scanblocks`): https://gist.github.com/jamesob/aa4a975344209f0316444b8de2ec1d18
### Original PR description
> The `scanblocks` RPC call allows one to get relevant blockhashes from a set of descriptors by scanning all blockfilters in a given range.
>
> **Example:**
>
> `scanblocks start '["addr(<bitcoin_address>)"]' 661000` (returns relevant blockhashes for `<bitcoin_address>` from blockrange 661000->tip)
>
> ## Why is this useful?
> **Fast wallet rescans**: get the relevant blocks and only rescan those via `rescanblockchain getblockheader(<hash>)[height] getblockheader(<hash>)[height])`. A future PR may add an option to allow to provide an array of blockhashes to `rescanblockchain`.
>
> **prune wallet rescans**: (_needs additional changes_): together with a call to fetch blocks from the p2p network if they have been pruned, it would allow to rescan wallets back to the genesis block in pruned mode (relevant #15946).
>
> **SPV mode** (_needs additional changes_): it would be possible to build the blockfilterindex from the p2p network (rather then deriving them from the blocks) and thus allow some sort of hybrid-SPV mode with moderate bandwidth consumption (related #9483)
ACKs for top commit:
furszy:
diff re-ACK 626b7c8
Tree-SHA512: f84e4dcb851b122b39e9700c58fbc31e899cdcf9b587df9505eaf1f45578cc4253e89ce2a45d1ff21bd213e31ddeedbbcad2c80810f46755b30acc17b07e2873
bf95976061 doc: add note about snapshot chainstate init (James O'Beirne)
e4d7995286 test: add testcases for snapshot initialization (James O'Beirne)
cced4e7336 test: move-only-ish: factor out LoadVerifyActivateChainstate() (James O'Beirne)
51fc9241c0 test: allow on-disk coins and block tree dbs in tests (James O'Beirne)
3c361391b8 test: add reset_chainstate parameter for snapshot unittests (James O'Beirne)
00b357c215 validation: add ResetChainstates() (James O'Beirne)
3a29dfbfb2 move-only: test: make snapshot chainstate setup reusable (James O'Beirne)
8153bd9247 blockmanager: avoid undefined behavior during FlushBlockFile (James O'Beirne)
ad67ff377c validation: remove snapshot datadirs upon validation failure (James O'Beirne)
34d1590331 add utilities for deleting on-disk leveldb data (James O'Beirne)
252abd1e8b init: add utxo snapshot detection (James O'Beirne)
f9f1735f13 validation: rename snapshot chainstate dir (James O'Beirne)
d14bebf100 db: add StoragePath to CDBWrapper/CCoinsViewDB (James O'Beirne)
Pull request description:
This is part of the [assumeutxo project](https://github.com/bitcoin/bitcoin/projects/11) (parent PR: https://github.com/bitcoin/bitcoin/pull/15606)
---
Half of the replacement for #24232. The original PR grew larger than expected throughout the review process.
This change adds the ability to initialize a snapshot-based chainstate during init if one is detected on disk. This is of course unused as of now (aside from in unittests) given that we haven't yet enabled actually loading snapshots.
Don't be scared! There are some big move-only commits in here.
Accompanying changes include:
- moving the snapshot coinsdb directory from being called `chainstate_[base blockhash]` to `chainstate_snapshot`, since we only support one snapshot in use at a time. This simplifies some logic, but it necessitates writing that base blockhash out to a file within the coinsdb dir. See [discussion here](https://github.com/bitcoin/bitcoin/pull/24232#discussion_r832762880).
- adding a simple fix in `FlushBlockFile()` that avoids a crash when attemping to flush to disk before `LoadBlockIndexDB()` is called, which happens when calling `MaybeRebalanceCaches()` during multiple chainstate init.
- improving the unittest to allow testing with on-disk chainstates - necessary to test a simulated restart and re-initialization.
ACKs for top commit:
naumenkogs:
utACK bf95976061
ariard:
Code Review ACK bf9597606
ryanofsky:
Code review ACK bf95976061. Changes since last review: rebasing, switching from CAutoFile to AutoFile, adding comments, switching from BOOST_CHECK to Assert in test util, using chainman.GetMutex() in tests, destroying one ChainstateManager before creating a new one in tests
fjahr:
utACK bf95976061
aureleoules:
ACK bf95976061
Tree-SHA512: 15ae75caf19f8d12a12d2647c52897904d27b265a7af6b4ae7b858592eeadb8f9da6c2394b6baebec90adc28742c053e3eb506119577dae7c1e722ebb3b7bcc0
861cb3fadc test: move SyncWithValidationInterfaceQueue() before Stop() in txindex_tests (Vasil Dimov)
6526dc3b78 test: silence TSAN false positive in coinstatsindex_initial_sync (Vasil Dimov)
Pull request description:
Silence false positives from TSAN about unsynchronized calls to `BaseIndex::~BaseIndex()` and `BaseIndex::SetBestBlockIndex()`. They are synchronized, but beyond the comprehension of TSAN - by `SyncWithValidationInterfaceQueue()`, called from `BaseIndex::BlockUntilSyncedToCurrentChain()`.
Fixes https://github.com/bitcoin/bitcoin/issues/25365
ACKs for top commit:
MarcoFalke:
review ACK 861cb3fadc
ryanofsky:
Code review ACK 861cb3fadc. Just comment change since last review.
Tree-SHA512: 8c30fdf2fd11d54e9adfa68a67185ab820bd7bd9f7f3ad6456e7e6d219fa9cf6d34b41e98e723eae86cb0c1baef7f3fc57b1b011a13dc3fe3d78334b9b5596de
b527b54950 net: convert standalone SetSocketNonBlocking() to Sock::SetNonBlocking() (Vasil Dimov)
29f66f7682 moveonly: move SetSocketNonBlocking() from netbase to util/sock (Vasil Dimov)
b4bac55679 net: convert standalone IsSelectableSocket() to Sock::IsSelectable() (Vasil Dimov)
5db7d2ca0a moveonly: move IsSelectableSocket() from compat.h to sock.{h,cpp} (Vasil Dimov)
Pull request description:
_This is a piece of #21878, chopped off to ease review._
* convert standalone `IsSelectableSocket()` to `Sock::IsSelectable()`
* convert standalone `SetSocketNonBlocking()` to `Sock::SetNonBlocking()`
This further encapsulates syscalls inside the `Sock` class and makes the callers mockable.
ACKs for top commit:
jonatack:
ACK b527b54950 review/debug build/unit tests at each commit, cross-referenced the changes with `man select` and `man errno`, ran a signet node on the last commit with ip4/ip6//tor/i2p/cjdns and network connections were nominal
dergoegge:
Code review ACK b527b54950
Tree-SHA512: af783ce558c7a89e173f7ab323fb3517103d765c19b5d14de29f64706b4e1fea3653492e8ea73ae972699986aaddf2ae72c7cfaa7dad7614254283083b7d2632
bcb0cacac2 reindex, log, test: fixes #21379 (mruddy)
Pull request description:
Fixes #21379.
The blocks/blk?????.dat files are mutated and become increasingly malformed, or corrupt, as a result of running the re-indexing process.
The mutations occur after the re-indexing process has finished, as new blocks are appended, but are a result of a re-indexing process miscalculation that lingers in the block manager's `m_blockfile_info` `nSize` data until node restart.
These additions to the blk files are non-fatal, but also not desirable.
That is, this is a form of data corruption that the reading code is lenient enough to process (it skips the extra bytes), but it adds some scary looking log messages as it encounters them.
The summary of the problem is that the re-index process double counts the size of the serialization header (magic message start bytes [4 bytes] + length [4 bytes] = 8 bytes) while calculating the blk data file size (both values already account for the serialization header's size, hence why it is over accounted).
This bug manifests itself in a few different ways, after re-indexing, when a new block from a peer is processed:
1. If the new block will not fit into the last blk file processed while re-indexing, while remaining under the 128MiB limit, then the blk file is flushed to disk and truncated to a size that is 8 greater than it should be. The truncation adds zero bytes (see `FlatFileSeq::Flush` and `TruncateFile`).
1. If the last blk file processed while re-indexing has logical space for the new block under the 128 MiB limit:
1. If the blk file was not already large enough to hold the new block, then the zeros are, in effect, added by `fseek` when the file is opened for writing. Eight zero bytes are added to the end of the last blk file just before the new block is written. This happens because the write offset is 8 too great due to the miscalculation. The result is 8 zero bytes between the end of the last block and the beginning of the next block's magic + length + block.
1. If the blk file was already large enough to hold the new block, then the current existing file contents remain in the 8 byte gap between the end of the last block and the beginning of the next block's magic + length + block. Commonly, when this occcurs, it is due to the blk file containing blocks that are not connected to the block tree during reindex and are thus left behind by the reindex process and later overwritten when new blocks are added. The orphaned blocks can be valid blocks, but due to the nature of concurrent block download, the parent may not have been retrieved and written by the time the node was previously shutdown.
ACKs for top commit:
LarryRuane:
tested code-review ACK bcb0cacac2
ryanofsky:
Code review ACK bcb0cacac2. This is a disturbing bug with an easy fix which seems well-worth merging.
mzumsande:
ACK bcb0cacac2 (reviewed code and did some testing, I agree that it fixes the bug).
w0xlt:
tACK bcb0cacac2
Tree-SHA512: acc97927ea712916506772550451136b0f1e5404e92df24cc05e405bb09eb6fe7c3011af3dd34a7723c3db17fda657ae85fa314387e43833791e9169c0febe51
04526787b5 Validate `port` options (amadeuszpawlik)
f8387c4234 Validate port value in `SplitHostPort` (amadeuszpawlik)
Pull request description:
Validate `port`-options, so that invalid values are rejected early in the startup.
Ports are `uint16_t`s, which effectively limits a port's value to <=65535. As discussed in https://github.com/bitcoin/bitcoin/pull/24116 and https://github.com/bitcoin/bitcoin/pull/24344, port "0" is considered invalid too.
Proposed in https://github.com/bitcoin/bitcoin/issues/21893#issuecomment-835784223
The `SplitHostPort(std::string in, uint16_t& portOut, std::string& hostOut)` now returns a bool that indicates whether the port value was set and within the allowed range. This is an improvement that can be used not only for port validation of options at startup, but also in rpc calls, etc,
ACKs for top commit:
luke-jr:
utACK 04526787b5
ryanofsky:
Code review ACK 04526787b5. Just suggested changes since last review: reverting some SplitHostPort changes, adding release notes, avoiding 'GetArgs[0]` problem.
Tree-SHA512: f1ac80bf98520b287a6413ceadb41bc3a93c491955de9b9319ee1298ac0ab982751905762a287e748997ead6198a8bb7a3bc8817ac9e3d2468e11ab4a0f8496d
43b8777dc3 refactor: move run_command from util to common (Cory Fields)
192325a77d kernel: move RunCommandParseJSON to its own file (Cory Fields)
Pull request description:
Because libbitcoinkernel does not include this new object, this has the side-effect of eliminating its unnecessary `boost::process` dependency.
This leaves libbitcoinkernel with 3 remaining boost dependencies:
- `boost::date_time` for `util/time.cpp`, which I'll separate out next. Exactly like this PR.
- `boost::signals2` for which I have a POC re-implementation here: https://github.com/theuni/bitcoin/commits/replace-boost-signals
- `boost::multi_index` which I'm not sure about yet.
ACKs for top commit:
ryanofsky:
Code review ACK 43b8777dc3. Could consider squashing the two commits, so the code just moves once instead of twice.
fanquake:
ACK 43b8777dc3
Tree-SHA512: f2a46cac34aaadfb8a1442316152ad354f6990021b82c78d80cae9fd43cd026209ffd62132eaa99d5d0f8cf34e996b6737d318a9d9a3f1d2ff8d17d697abf26d
faa15527d7 test: Use dedicated mempool in TestBasicMining (MacroFake)
fafab384a0 test: Use dedicated mempool in TestPackageSelection (MacroFake)
fa4055d79c test: Use dedicated mempool in TestPrioritisedMining (MacroFake)
fa29218285 test: Pass mempool reference to AssemblerForTest (MacroFake)
Pull request description:
This cleans up the miner tests:
* Removes duplicate/redundant and thus confusing chainparams object.
* Uses a fresh mempool for each subtest instead of using the "global" one from the testing setup. This makes it easier to follow the tests in smaller scopes. Also it makes sure the mempool is truly cleared by reconstructing it. Finally, this removes calls to `clear`, see https://github.com/bitcoin/bitcoin/pull/19909
ACKs for top commit:
glozow:
utACK faa15527d7
Tree-SHA512: ced1260f6ab70fba74b0fac7ff4fc7adfddcd2f3bee785249d2a4a9055ac253eff9090edbda7a17e72a71a81b56ff708d5ff64e1f57ebc7b7747d6c88fec51e3
33b12e5df6 docs: improve docs where MemPoolLimits is used (stickies-v)
6945853c0b test: use NoLimits() in MempoolIndexingTest (stickies-v)
3a86f24a4c refactor: mempool: use CTxMempool::Limits (stickies-v)
b85af25f87 refactor: mempool: add MemPoolLimits::NoLimits() (stickies-v)
Pull request description:
Mempool currently considers 4 limits regarding ancestor and descendant count and size, which get passed around between functions quite a bit. This PR uses `CTxMemPool::Limits` introduced in https://github.com/bitcoin/bitcoin/pull/25290 to simplify those signatures and callsites.
The purpose of this PR is to improve readability and maintenance, without behaviour change.
As noted in the first commit "refactor: mempool: change MemPoolLimits members to uint", we currently have an underflow issue where a user could pass a negative `-limitancestorsize`, which is eventually cast to an unsigned integer. This behaviour already exists. Because it's orthogonal and to minimize scope, I think this should be fixed in a separate PR.
ACKs for top commit:
hebasto:
ACK 33b12e5df6, I have reviewed the code and it looks OK, I agree it can be merged.
glozow:
reACK 33b12e5df6
Tree-SHA512: 591c6dcee1894f1c3ca28b34a680eeadcf0d40cda92451b4a422c03087b27d682b5e30ba4367abd75a99b5ccb115b7884b0026958d3c7dddab030549db5a4056
Forward the validation of the port from `ParseUInt16(...)`.
Consider port 0 as invalid.
Add suitable test for the `SplitHostPort` function.
Add doxygen description to the `SplitHostPort` function.
The (100, 1000000, 1000, 1000000) limits are arbitrarily high and
don't restrict anything, they are just meant to calculate ancestors
properly. Using NoLimits() makes this intent more clear and simplifies
the code.
8a6b6dfcd8 fuzz: pass max fee into ConsumeTxMemPoolEntry (fanquake)
eb15569280 fuzz: add util/mempool/h.cpp (fanquake)
Pull request description:
Moving the heavy (Boost) mempool code out of fuzz/util.h. Means that (for ex) a crypto_common fuzz unit doesn't need to care about seeing endless Boost headers. This results in a ~10% speedup (for me) when compiling the fuzz tests. Your results may vary.
ACKs for top commit:
MarcoFalke:
review ACK 8a6b6dfcd8🍮
Tree-SHA512: 27dc9d9581ac0b1b319cc0dc08fe5f8fbf9269386a5cb23f6fd5d8231bf015ed942ab4414d8001220541be0013756354578ddab1fec607c6fba04daf421bc870
Quoting ryanofsky: "util can be the library for things included in the kernel
which the kernel can depend on, and common can be the library for other code
that needs to be shared internally, but should not be part of the kernel or
shared externally."
079cf88c0d refactor: move Boost datetime usage to wallet (fanquake)
Pull request description:
This means we don't need Boost Datetime in a `--disable-wallet` build, and it isn't included in the kernel (via time.h/cpp). Split from a larger boost removal branch/effort.
ACKs for top commit:
hebasto:
re-ACK 079cf88c0d
aureleoules:
re-ACK 079cf88c0d - rebased and two additional unit tests since my last review.
jarolrod:
crACK 079cf88c0d
Tree-SHA512: c84f47158a4f21902f211c059d8c4bd55ffe95a256835deee723653be08cca49eeddfc33a2316b0cd31805e81cf77eaa39c6c9dcff4cda11a26ba4c1c143974e
