c318211ddd walletdb: fix last client version update (furszy)
bda8ebe608 wallet: don't read db every time that a new WalletBatch is created (furszy)
Pull request description:
Found it while was working on #25297.
We are performing a db read operation every time that a new `WalletBatch` is created, inside the constructor, just to check if the client version field is inside the db or not.
As the client version field does not change in the entire db lifecycle, this operation can be done only once: The first time that the db is accessed/opened and the client version value can be cached.
ACKs for top commit:
achow101:
ACK c318211ddd
w0xlt:
reACK c318211ddd
Tree-SHA512: 7fb780c656e169e8eb21e7212242494a647f6506d6da2cca828703713d440d29c82bec9e7d2c410f37b49361226ccd80846d3eeb8168383d0c2a11d85d73bee2
25e4762ae7 wallet: more accurate tx_noinputs_size (S3RK)
Pull request description:
Rationale: more accurate non-input fee estimation for txs with >=253 inputs
ACKs for top commit:
laanwj:
Concept and code review ACK 25e4762ae7
achow101:
ACK 25e4762ae7
furszy:
Code review ACK 25e4762a. left a small nit.
Tree-SHA512: bda8fad725d32ad3e13c007fa56ddb6679ac1a32098ddb08d9a114054acfa681cb66cd703ac675297f731cb381b09067a99a4efa31320140bbdd03f0cfdc81af
This reverts commit 9b5950db86.
Waste can be negative. At feerates lower than long_term_feerate this
means that a waste of 0 may be a suboptimal solution and this causes the
search to exit prematurely.
Only when the feerate is equal to the long_term_feerate would achieving
a waste of 0 indicate that we have achieved an optimal solution,
because it would mean that the excess is 0. It seems unlikely
that this would ever occur outside of test cases, and even then we
should prefer solutions with more inputs over solutions with fewer
according to previous decisions—but solutions with more inputs are found
later in the branch exploration.
The "optimization" described in #18257 and implemented in #18262 is
therefore a premature exit on a suboptimal solution and should be reverted.
e673d8b475 bench: Enable loading benchmarks depending on what's compiled (Andrew Chow)
4af3547eba bench: Use mock wallet database for wallet loading benchmark (Andrew Chow)
49910f255f sqlite: Use in-memory db instead of temp for mockdb (Andrew Chow)
a1080802f8 walletdb: Create a mock database of specific type (Andrew Chow)
7c0d34476d bench: reduce the number of txs in wallet for wallet loading bench (Andrew Chow)
f85b54ed27 bench: Add transactions directly instead of mining blocks (Andrew Chow)
d94244c4bf bench: reduce number of epochs for wallet loading benchmark (Andrew Chow)
817c051364 bench: use unsafesqlitesync in wallet loading benchmark (Andrew Chow)
9e404a9831 bench: Remove minEpochIterations from wallet loading benchmark (Andrew Chow)
Pull request description:
`minEpochIterations` is probably unnecessary to set, so removing it makes the runtime much faster.
ACKs for top commit:
Rspigler:
tACK e673d8b475
furszy:
Code review ACK e673d8b4, nice PR.
glozow:
Concept ACK e673d8b475. For each commit, verified that there was a performance improvement without negating the purpose of the bench, and made some effort to verify that the code is correct.
Tree-SHA512: 9337352ef846cf18642d5c14546c5abc1674b4975adb5dc961a1a276ca91f046b83b7a5e27ea6cd26264b96ae71151e14055579baf36afae7692ef4029800877
When filling a PSBT, we search the listed pubkeys in order to determine
whether the current DescriptorScriptPubKeyMan could sign the transaction
even if it is not watching the scripts. With Taproot, the taproot
pubkeys need to be searched as well.
Currently, the wallet scan progress is not saved.
If it is interrupted, it will be necessary to start from
scratch on the next load.
With this change, progress is saved every 60 seconds.
Co-authored-by: furszy <matiasfurszyfer@protonmail.com>
Co-authored-by: Jon Atack <jon@atack.com>
Co-authored-by: Ryan Ofsky <ryan@ofsky.org>
32e5edc0f4 wallet: avoid extra wtx lookup in AddToSpends (furszy)
Pull request description:
As `AddToSpends` is only called from `AddToWallet` and `LoadToWallet`, places where we insert the wtx into the wallet map, we can directly feed `AddToSpends` with the `wtx` and remove another extra lookup.
ACKs for top commit:
laanwj:
Code review ACK 32e5edc0f4
achow101:
ACK 32e5edc0f4
theStack:
Code-review ACK 32e5edc0f4
w0xlt:
Code Review ACK 32e5edc0f4
brunoerg:
crACK 32e5edc0f4
Tree-SHA512: e9fb8df44c3e3fa26c107d261bf78e45014b4755890a64817f2be62ee6b7751f5dd2813a18dcb103a21ddba1422f9d2d59c4bf186f08314e634365d36b01be8f
Otherwise, RPC commands such as `walletcreatefundedpsbt` will not support the manual selection of locked, spent and externally added coins.
Full explanation is inside #25118 comments but brief summary is:
`vCoins` at `SelectCoins` time could not be containing the manually selected input because, even when they were selected by the user, the current `AvailableCoins` flow skips locked and spent coins.
Extra note: this is an intermediate step to unify the `fAllowOtherInputs`/`m_add_inputs` concepts. It will not be a problem anymore in the future when we finally decouple the wtx-outputs lookup process from `SelectCoins` and don't skip the user's manually selected coins in `AvailableCoins`.
Seeking to make the `CoinControl` option less confusing/redundant.
In #16377 the `CoinControl` flag ‘m_add_inputs’ was added to tell the coin filtering and selection process two things:
- Coin Filtering: Only use the provided inputs. Skip the Rest.
- Coin Selection: Search the wtxs-outputs and append all the `CoinControl` selected outpoints to the selection result (skipping all the available output checks). Nothing else.
Meanwhile, in `CoinControl` we already have a flag ‘fAllowOtherInputs’ which is already saying:
- Coin Filtering: Only use the provided inputs. Skip the Rest.
- Coin Selection: If false, no selection process -> append all the `CoinControl` selected outpoints to the selection result (while they passed all the `AvailableCoins` checks and are available in the 'vCoins' vector).
As can notice, the first point in the coin filtering process is duplicated in the two option flags. And the second one, is slightly different merely because it takes into account whether the coin is on the `AvailableCoins` vector or not.
So it makes sense to merge ‘m_add_inputs’ and ‘fAllowOtherInputs’ into a single field for the coin filtering process while introduce other changes to add the missing/skipped coins into 'vCoins' vector if they were manually selected by the user (follow-up commits).
fd5c996d16 wallet: GetAvailableBalance, remove double walk-through every available coin (furszy)
162d4ad10f wallet: add 'only_spendable' filter to AvailableCoins (furszy)
cdf185ccfb wallet: remove unused IsSpentKey(hash, index) method (furszy)
4b83bf8dbc wallet: avoid extra IsSpentKey -> GetWalletTx lookups (furszy)
3d8a282257 wallet: decouple IsSpentKey(scriptPubKey) from IsSpentKey(hash, n) (furszy)
a06fa94ff8 wallet: IsSpent, 'COutPoint' arg instead of (hash, index) (furszy)
91902b7720 wallet: IsLockedCoin, 'COutPoint' arg instead of (hash, index) (furszy)
9472ca0a65 wallet: AvailableCoins, don't call 'wtx.tx->vout[i]' multiple times (furszy)
4ce235ef8f wallet: return 'CoinsResult' struct in `AvailableCoins` (furszy)
Pull request description:
This started in #24845 but grew out of scope of it.
So, points tackled:
1) Avoid extra `GetWalletTx` lookups inside `AvailableCoins -> IsSpentKey`.
`IsSpentKey` was receiving the tx hash and index to internally lookup the tx inside the wallet's map. As all the `IsSpentKey` function callers already have the wtx available, them can provide the `scriptPubKey` directly.
2) Most of the time, we call `Wallet::AvailableCoins`, and later on the process, skip the non-spendable coins from the result in subsequent for-loops. So to speedup the process: introduced the ability to filter by "only_spendable" coins inside `Wallet::AvailableCoins` directly.
(the non-spendable coins skip examples are inside `AttemptSelection->GroupOutputs` and `GetAvailableBalance`).
4) Refactored `AvailableCoins` in several ways:
a) Now it will return a new struct `CoinsResult` instead of receiving the vCoins vector reference (which was being cleared at the beginning of the method anyway). --> this is coming from #24845 but cherry-picked it here too to make the following commits look nicer.
b) Unified all the 'wtx.tx->vout[I]' calls into a single call (coming from this comment https://github.com/bitcoin/bitcoin/pull/24699#discussion_r854163032).
5) The wallet `IsLockedCoin` and `IsSpent` methods now accept an `OutPoint` instead of a hash:index. Which let me cleanup a bunch of extra code.
6) Speeded up the wallet 'GetAvailableBalance': filtering `AvailableCoins` by spendable outputs only and using the 'AvailableCoins' retrieved `total_amount` instead of looping over all the retrieved coins once more.
-------------------------------------------------------
Side topic, all this process will look even nicer with #25218
ACKs for top commit:
achow101:
ACK fd5c996d16
brunoerg:
crACK fd5c996d16
w0xlt:
Code Review ACK fd5c996d16
Tree-SHA512: 376a85476f907f4f7d1fc3de74b3dbe159b8cc24687374d8739711ad202ea07a33e86f4e66dece836da3ae6985147119fe584f6e672f11d0450ba6bd165b3220
The value was only being updated launching releases with higher version numbers
and not if the user launched a previous release.
Co-authored-by: MacroFake <falke.marco@gmail.com>
7832e9438f test: fundrawtransaction preset input weight calculation (S3RK)
c3981e379f wallet: do not count wallet utxos as external (S3RK)
Pull request description:
Correctly differentiating between external vs non-external utxos in coin control produces more accurate weight and fee estimations.
Weight for external utxos is estimated based on the maximum signature size, while for the wallet utxos we expect minimal signature due to signature grinding.
ACKs for top commit:
achow101:
re-ACK 7832e9438f
Xekyo:
re-ACK 7832e9438f
furszy:
ACK 7832e943
Tree-SHA512: bb5635b0bd85fa9a76922a53ad3fa062286424c06a695a0e87407c665713e80a33555b644fbb13bcc1ab503dcd7f53aacbdc368d69ac0ecff8005603623ac94f
e3609cdc01 doc: Update importaddress mention incompatibility with descriptor wallet (BrokenProgrammer)
Pull request description:
This is related to #25363 and offers a small update to the error messages from `EnsureLegacyScriptPubKeyMan` and `EnsureConstLegacyScriptPubKeyMan` to mention that they only are compatible with legacy wallets.
The RPC documentation for `importaddress` is also updated to mention this as well as guide the user to the alternative `importdescriptors` for cases when using descriptor wallets.
I'm thinking that we can introduce a "porting guide" document mentioned in #25363 in a separate PR since I would have to make myself more familiar with the subject before being able to tackle that.
ACKs for top commit:
laanwj:
Code review ACK e3609cdc01
achow101:
ACK e3609cdc01
Tree-SHA512: c7a924a7283fe59dc4e04c8c8fa034c15601f0b25eff09d975e98e2e8db5268ff470336b2d978d6916af9f782f9257b840d64bd15485b1742b4a8b8bfd0bb50f
ce893c0497 doc: Update developer notes (Anthony Towns)
d2852917ee sync.h: Imply negative assertions when calling LOCK (Anthony Towns)
bba87c0553 scripted-diff: Convert global Mutexes to GlobalMutexes (Anthony Towns)
a559509a0b sync.h: Add GlobalMutex type (Anthony Towns)
be6aa72f9f qt/clientmodel: thread safety annotation for m_cached_tip_mutex (Anthony Towns)
f24bd45b37 net_processing: thread safety annotation for m_tx_relay_mutex (Anthony Towns)
Pull request description:
This changes `LOCK(mutex)` for non-global, non-recursive mutexes to be annotated with the negative capability for the mutex it refers to, to prevent . clang applies negative capabilities recursively, so this helps avoid forgetting to annotate functions.
This can't reasonably be used for globals, because clang would require every function to be annotated with `EXCLUSIVE_LOCKS_REQUIRED(!g_mutex)` for each global mutex; so this introduces a trivial `GlobalMutex` subclass of `Mutex`, and reduces the annotations for both `GlobalMutex` to `LOCKS_EXCLUDED` which only catches trivial errors (eg (`LOCK(x); LOCK(x);`).
ACKs for top commit:
MarcoFalke:
review ACK ce893c0497🐦
hebasto:
ACK ce893c0497
Tree-SHA512: 5c35e8c7677ce3d994a7e3774f4344adad496223a51b3a1d1d3b5f20684b2e1d5cff688eb3fbc8d33e1b9940dfa76e515f9434e21de6f3ce3c935e29a319f529
If there are multiple external signers, `GetExternalSigner()` will
just pick the first one in the list. If the user has two or more
hardware wallets connected at the same time, he might not notice this.
This PR adds a check and fails with suitable message.
We are skipping the non-spendable coins that appear in vCoins ('AvailableCoins' result) later, in several parts of the CreateTransaction and GetBalance flows:
GetAvailableBalance (1) gets all the available coins calling AvailableCoins and, right away, walk through the entire vector, skipping the non-spendable coins, to calculate the total balance.
Inside CreateTransactionInternal —> SelectCoins(vCoins,...), we have several calls to AttemptSelection which, on each of them internally, we call twice to GroupOutputs which internally has two for-loops over the entire vCoins vector that skip the non-spendable coins.
So, Purpose is not add the non-spendable coins into the AvailableCoins result (vCoins) in the first place for the processes that aren’t using them at all, so we don’t waste resources skipping them later so many times.
Note: this speedup is for all the processes that call to CreateTransaction and GetBalance* internally.
This will be used in a follow-up commit to prevent extra 'GetWalletTx' lookups if the function caller already have the wtx and can just provide the scriptPubKey directly.
Instead of accepting a `vCoins` reference that is cleared at the beginning of the method.
Note:
This new struct, down the commits line, will contain other `AvailableCoins` useful results.
57fb37c275 wallet: CommitTransaction, remove extra wtx lookup and add exception for a possible db write error. (furszy)
Pull request description:
Two points for `CWallet::CommitTransaction`:
1) The extra wtx lookup:
As we are calling to `AddToWallet` first, which returns the recently added/updated wtx pointer, there is no need to look up the wtx again few lines later. We can just use it.
2) The db write error:
`AddToWallet` can only return a nullptr if the db write fails, which inside `CommitTransaction` translates to an exception throw cause. We expect everywhere that `CommitTransaction` always succeed.
------------------------------------------------
Extra note:
This finding generated another working path for me :)
It starts with the following question: why are we returning a nullptr from `AddToWallet` if the db write failed without removing the recently added transaction from the wallet's map?..
Can led to a wallet invalid state where the inputs of this new transaction are not marked dirty, while the transaction that spends them still exist on the in-memory wallet tx map.
-- I'm writing it here to gather some feedback first and not forget it, will create a follow-up PR in the coming days 🚜 --
ACKs for top commit:
achow101:
ACK 57fb37c275
jonatack:
ACK 57fb37c
ryanofsky:
Code review ACK 57fb37c275. Seems like a clear improvement. Better to fail earlier with a better error message if the failure is going to happen anyway
Tree-SHA512: 80e59c01852cfbbc70a5de1a1c2c59b5e572f9eaa08c2175112cb515256e63fa04c7942f92a513b620d6b06e66392029ebe8902287c456efdbee58a7a5ae42da
3a9b9bb38e test: ensure createmultisig and addmultisigaddress are not returning any warning for expected cases (brunoerg)
eaf6f630c0 rpc: fix inappropriate warning for address type p2sh-segwit in createmultisig and addmultisigaddress (brunoerg)
Pull request description:
Fixes#25127
If there are any uncompressed keys when calling `AddAndGetMultisigDestination`, it will just default to a legacy address regardless of the chosen `address_type`. So, #23113 added a warnings field which will warn the user why their address format is different.
However, when creating a multisig (p2sh-segwit), it is returning an inappropriate warning, because when getting the output type from destination (`OutputTypeFromDestination`), it returns `ScriptHash` for both legacy and `P2SH_SEGWIT`. So, since `P2SH_SEGWIT` is different from `ScriptHash`, it returns the warning:
192d639a6b/src/rpc/output_script.cpp (L166-L169)
So, to avoid this mistake I changed `OutputTypeFromDestination` to `descriptor->GetOutputType()` to get the appropriate output type.
ACKs for top commit:
jonatack:
ACK 3a9b9bb38e
laanwj:
Code review ACK 3a9b9bb38e
Tree-SHA512: 49f717479c2b8906277e7591ddd4747f7961c2d5c77494b5124045de9036a4277d46b9ad99279d51f0c4484284c445f1e1d3c55c49bbf0716741bad426a89369
