...also adjust callers
Changes:
- In BlockAssembler::CreateNewBlock, we now only lock m_mempool->cs and
call addPackageTxs if m_mempool is not nullptr
- BlockAssembler::addPackageTxs now takes in a mempool reference, and is
annotated to require that mempool's lock.
- In TestChain100Setup::CreateBlock and generateblock, don't construct
an empty mempool, just pass in a nullptr for mempool
Since UpdatePackagesForAdded is a helper function that's only used in
addPackageTxs we can make it static and avoid the unnecessary interface
and in-header lock annotation.
SkipMapTxEntry is a short helper function that's only used in
addPackageTxs, we can just inline it, keep the comments, and avoid the
unnecessary interface and lock annotations.
7036cf52aa Delete UpdatePackagesForAdded at beginning of addPackageTxs. (KevinMusgrave)
Pull request description:
In `CreateNewBlock` (in miner.cpp), `inBlock` is cleared before `addPackageTxs`, so `inBlock` will be empty in the first call to `UpdatePackagesForAdded`. I saw this brought up in these [PR review club logs](https://bitcoincore.reviews/24538) and there didn't seem to be a definitive answer for why the call is necessary. There's also an [old PR](https://github.com/bitcoin/bitcoin/pull/10200) where this change was going to be applied, but it got closed.
If `addPackageTxs` can be called when `inBlock` is not empty, then maybe a test should be added for that case. All the tests seem to pass with this deletion.
ACKs for top commit:
glozow:
utACK 7036cf52aa
Tree-SHA512: 9e757b71b9035f68a0c6fef229b8cd83f1bdbe23f05bb02cc1bab8c3c177805b388bceb2bb1f0bce354791ccb29f351a6c51979b96ffe4d9fc6c978f83e36afc
f9fdcec7e9 settings: Add resetSettings() method (Ryan Ofsky)
77fabffef4 init: Remove Shutdown() node.args reset (Ryan Ofsky)
0e55bc6e7f settings: Add update/getPersistent/isIgnored methods (Ryan Ofsky)
Pull request description:
Add `interfaces::Node` `updateSetting`, `forceSetting`, `resetSettings`, `isSettingIgnored`, and `getPersistentSetting` methods so GUI is able to manipulate `settings.json` file and use and modify node settings.
(Originally this PR also contained GUI changes to unify bitcoin-qt and bitcoind persistent settings and call these methods, but the GUI commits have been dropped from this PR and moved to bitcoin-core/gui/pull/602)
ACKs for top commit:
vasild:
ACK f9fdcec7e9
hebasto:
re-ACK f9fdcec7e9, only a function renamed since my recent [review](https://github.com/bitcoin/bitcoin/pull/15936#pullrequestreview-979324357).
Tree-SHA512: 4cac853ee29be96d2ff38404165b9dfb7c622b2a9c99a15979596f3484ffde0da3d9c9c372677dff5119ca7cffa6383d81037fd9889a29cc9285882a8dc0c268
664a14ba7c coinstats: Move GetUTXOStats to rpc/blockchain (Carl Dong)
f100687566 kernel: Use ComputeUTXOStats in validation (Carl Dong)
faa52387e8 style-only: Rearrange using decls after scripted-diff (Carl Dong)
f329a9298c scripted-diff: Move src/kernel/coinstats to kernel:: (Carl Dong)
0e54456f04 Use only kernel/coinstats.h in index/coinstatsindex.h (Carl Dong)
80970985c9 coinstats: Split node/coinstats.h to kernel/coinstats.h (Carl Dong)
35f73ce4b2 coinstats: Move hasher codepath to kernel/coinstats (Carl Dong)
b7634fe02b Move logic from LookupUTXOStatsWithIndex to CoinStatsIndex::LookUpStats (Carl Dong)
1352e410a5 coinstats: Separate hasher/index lookup codepaths (Carl Dong)
524463daf6 coinstats: Return purely out-param CCoinsStats (Carl Dong)
46eb9fc56a coinstats: Extract index_requested in-member to in-param (Carl Dong)
a789f3f2b8 coinstats: Extract hash_type in-member to in-param (Carl Dong)
102294898d includes: Remove rpc/util.h -> node/coinstats.h (Carl Dong)
0848db9c35 fuzz: Remove useless GetUTXOStats fuzz case (Carl Dong)
52b1939993 kernel: Remove unnecessary blockfilter{index,}.cpp (Carl Dong)
Pull request description:
Part of: #24303
Depends on: #24322
The `GetUTXOStats` function has 2 codepaths:
- One which queries the `CoinStatsIndex` for the UTXO hash
- One which actually performs the hashing
For `libbitcoinkernel`, the only place where we call `GetUTXOStats` is in `PopulateAndValidateSnapshots`, which uses the `SHA256D` hash, and is therefore unable to use the `CoinStatsIndex` since that only provides `MuHash` hashes. Not that I think indices necessarily belong in `libbitcoinkernel` anyway.
This PR separates these 2 aforementioned codepaths of `GetUTXOStats`, uses the hashing codepath in `PopulateAndValidateSnapshots`, and removes the need to link in `index/coinstatsindex.cpp` and `node/coinstats.cpp`.
-----
Logistically, this PR:
- Extracts out the `index_requested` and `hash_type` members of `CoinStats`, which served as "in-params" to `GetUTXOStats` embedded within the `CoinStats` struct. This allows `CoinStats` to only consist of "out-param" members, and be returned by `GetUTXOStats` without needing to be an "in-out" param
- Introduce the purely virtual `UTXOHashers` class, with 3 implementations: `SHA256DHasher`, `MuHashHasher`, and `NullHasher`. These replace the existing template-based polymorphism.
- Split `GetUTXOStats` into:
- `CalculateUTXOStatsWithHasher(UTXOHasher&, ...)`, and
- `LookupUTXOStatsWithIndex(CoinStatsIndex&, ...)`
- Use `CalculateUTXOStatsWithHasher` directly where appropriate (`src/validation.cpp` and `src/fuzz`)
- Move `GetUTXOStats` to `rpc/blockchain`, which is the only place that depends on `GetUTXOStats`'s weird fallback behaviour
- Move `LookupUTXOStatsWithIndex` to `index/coinstatsindex`
Code organization:
- `src/`
- `kernel/` → only contains the hashing codepath
- `coinstats.cpp` → hashing codepath implementations
- `coinstats.h` → header for `kernel/coinstats.cpp`
- `index/` → only contains the index codepath
- `coinstatsindex.cpp` → index codepath implementations
- `coinstatsindex.h`
- `validation.cpp` → only uses the hashing codepath
- `rpc/blockchain.cpp` → uses both the hashing and index codepath, old `GetUTXOStats` fallback logic moved here as static
- `test/fuzz/coins_view.cpp` → only uses the hashing codepath
TODOs:
- [x] Commit messages could be fleshed out more
Would love any feedback!
ACKs for top commit:
laanwj:
Code review ACK 664a14ba7c
Tree-SHA512: 18722c7bd279174d2d1881fec33ea04a9b261aae1c12e998cf434ef297d8ded47de69c526c8033a2ba7abc93ba3d2ff5faf4ce05e8888c725c31cf885ce3ef73
rpc/blockchain.cpp is now the only user of the vestigial
GetUTXOStats(...). And since GetUTXOStats(...)'s special fallback logic
was only really relevant/meant for rpc/blockchain.cpp, we can just move
it there.
Most of this commit is pure-move.
After this change:
- kernel/coinstats.h
-> Contains declarations for:
- enum class CoinStatsHashType
- struct CCoinsStats
- GetBogoSize(...)
- TxOutSer(...)
- ComputeUTXOStats(...)
- node/coinstats.h
-> Just GetUTXOStats, which will be removed as we change callers to
directly use the hashing/indexing codepaths in future commits.
As mentioned in a previous commit, the hashing codepath can now be moved
to a separate file. This decouples callers that only rely on the hashing
codepath from the indexing one.
This is key for libbitcoinkernel, which needs to have the CoinsStats
hashing codepath for AssumeUTXO, but does not wish to be coupled with
indexes.
Note that only the .cpp file is split in this commit, the header files
will be split in a subsequent commit and the #includes to
node/coinstats.h will be adjusted to only #include the necessary
headers.
The indexing codepath logic in node/coinstats.cpp is simple enough to be
moved into CoinStatsIndex::LookUpStats, avoiding an additional layer of
function calls. Callers are modified accordingly.
Also, add 2 missed BOOST_CHECKs to the coinstatsindex_initial_sync unit
test.
Split out ComputeUTXOStats and LookupUTXOStatsWithIndex from
GetUTXOStats, since the hashing and indexing codepaths are quite
disparate in practice.
Also allow add a constructor to CCoinsStats for it to be constructed
from a a block height and hash. This is used in both codepaths.
Also add a note in GetUTXOStats documenting a behaviour quirk that
predates this patchset.
[META] This allows the hashing codepath to be moved to a separate file
in a future commit, decoupling callers that only rely on the
hashing codepath from the indexing one. This is key for
libbitcoinkernel, which needs to have the hashing codepath for
AssumeUTXO, but does not wish to be coupled with indexes.
In previous commits in this patchset, we removed all in-param members of
CCoinsStats. Now that that's done, we can modify GetUTXOStats to return
an optional CCoinsStats instead of a status bool. Callers are modified
accordingly.
In rpc/blockchain.cpp, we discover that GetUTXOStats' status bool when
getting UTXO stats for pprev was not checked for error. We fix this as
well.
Since the removal of NODISCARD in 81d5af42f4,
the only attributes def is LIFETIMEBOUND, and it's included in many more
places that it is used.
This removes all includes which do not have an associated use of LIFETIMEBOUND,
and adds it to the following files, due to their use of the same:
* src/validationinterface.h
* src/script/standard.h
Currently, CCoinsStats is a struct with both in-params and out-params
where the hash_type and index_requested members are the only in-params.
This change removes CCoinsStats' hash_type in-param member and adds it
to the relevant functions instead.
[META] In subsequent commits, all of CCoinsStats' members which serve as
in-params will be moved out so as to make CCoinsStats a pure
out-param struct.
This decouples validation.cpp from netaddress.cpp (transitively,
timedata.cpp, and asmap.cpp).
This is important for libbitcoinkernel as:
- There is no reason for the consensus engine to be coupled with
netaddress, timedata, and asmap
- Users of libbitcoinkernel can now easily supply their own
std::function that provides the adjusted time.
See the src/Makefile.am changes for some satisfying removals.
fa1b76aeb0 Do not call global Params() when chainman is in scope (MacroFake)
fa30234be8 Do not pass CChainParams& to PeerManager::make (MacroFake)
fafe5c0ca2 Do not pass CChainParams& to BlockAssembler constructor (MacroFake)
faf012b438 Do not pass Consensus::Params& to Chainstate helpers (MacroFake)
fa4ee53dca Do not pass time getter to Chainstate helpers (MacroFake)
Pull request description:
It seems confusing to pass chain params, consensus params, or a time function around when it is not needed.
Fix this by:
* Inlining the passed time getter function. I don't see a use case why this should be mockable.
* Using `chainman.GetConsensus()` or `chainman.GetParams()`, where possible.
ACKs for top commit:
promag:
Code review ACK fa1b76aeb0.
vincenzopalazzo:
ACK fa1b76aeb0
Tree-SHA512: 1abff5cba4b4871d97f17dbcdf67bc9255ff21fa4150a79a74e39b28f0610eab3e7dee24d56872dd6e111f003b55e288958cdd467e6218368d896f191e4ec9cd
Allows the GUI to clear settings.json file and save settings.json.bak file when
GUI "Reset Options" button is pressed or -resetguisettings command line option
is used. (GUI code already backs up and resets the "guisettings.ini" file this
way, so this just makes the same behavior possible for "settings.json")
fa9af21878 scripted-diff: Use getInt<T> over get_int/get_int64 (MacroFake)
Pull request description:
Seems better to see the return type directly and be able to modify it easier, as the return type is used for exceptions (in-range checking and parsing feedback).
ACKs for top commit:
fanquake:
ACK fa9af21878
Tree-SHA512: 284aa2527d0f663ca01550115025c9c64c787531d595f866c718f6ad09b9b0cac1e683a7d77f8009b75de990fd37166b44063ffa83fba8a04e9a31600b4c2725
Add interfaces::Node methods to give GUI finer grained control over
settings.json file. Update method is used to write settings to the file,
getPersistent and isIgnored methods are used to find out about settings
file and command line option interactions.
e4303c337c [unit test] prioritisation in mining (glozow)
7a8d60676b [miner] bug fix: update for parent inclusion using modified fee (glozow)
0f9a44461c MOVEONLY: group miner tests into MinerTestingSetup functions (glozow)
Pull request description:
Came up while reviewing #24364, where some of us incorrectly assumed that we use the same fee deduction in `CTxMemPoolModifiedEntry::nModFeesWithAncestors` when first constructing an entry and in `update_for_parent_inclusion`.
Actually, the behavior is this: when a mempool entry's ancestor is included in the block template, we create a `CTxMemPoolModifiedEntry` for it, subtracting the ancestor's modified fees from `nModFeesWithAncestors`. If another ancestor is included, we update it again, but use the ancestor's _base_ fees instead.
I can't explain why we use `GetFee` in one place and `GetModifiedFee` in the other, but I'm quite certain we should be using the same one for both.
And should it be base or modified fees? Modified, otherwise the child inherits the prioritisation of the parent, but only until the parent gets mined. If we want prioritisation to cascade down to current in-mempool descendants, we should probably document that in the `prioritsetransaction` helpstring and implement it in `CTxMemPool::mapDeltas`, not as a quirk in the mining code?
Wrote a test in which a mempool entry has 2 ancestors, both prioritised, and both included in a block template individually. This test should fail without the s/GetFee/GetModifiedFee commit.
ACKs for top commit:
ccdle12:
tested ACK e4303c3
MarcoFalke:
ACK e4303c337c🚗
Tree-SHA512: 4cd94106fbc9353e9f9b6d5af268ecda5aec7539245298c940ca220606dd0737264505bfaae1f83d94765cc2d9e1a6e913a765048fe6c19292482241761a6762
In previous commits in this patchset, we've made sure that every
Unload/UnloadBlockIndex member function resets its own members, and does
not reach out to globals.
This means that their corresponding classes' default destructors can now
replace them, and do an even more thorough job without the need to be
updated for every new member variable.
Therefore, we can remove them, and also remove UnloadBlockIndex since
that's not used anymore.
Unfortunately, chainstatemanager_loadblockindex relies on
CChainState::UnloadBlockIndex, so that needs to stay for now.
Fixes https://github.com/bitcoin/bitcoin/issues/22964
Previously, we used UnloadBlockIndex() in order to reset node.mempool
and node.chainman. However, that has proven to be fragile (see
https://github.com/bitcoin/bitcoin/issues/22964), and requires
UnloadBlockIndex and its callees to be updated manually for each member
that's introduced to the mempool and chainman classes.
In this commit, we stop using the UnloadBlockIndex function and we
simply reconstruct node.mempool and node.chainman.
Since PeerManager needs a valid reference to both node.mempool and
node.chainman, we also move PeerManager's construction via `::make` to
after the chainstate activation sequence is complete.
There are no more callers to UnloadBlockIndex after this commit, so it
and its sole callees can be pruned.
fa1970f075 Make BlockManager::LoadBlockIndex private (MarcoFalke)
Pull request description:
* After commit fa27f03b49 `BlockManager::LoadBlockIndex` is only called by `BlockManager::LoadBlockIndexDB`. Thus, it can be made `private`.
* After commit c600ee3816 `m_best_invalid` is no longer accessed by `BlockManager::LoadBlockIndex`. Thus, the unused `friend` can be removed.
ACKs for top commit:
mruddy:
ACK fa1970f075 I verified by double checking references, then applying the patch, and running `make check`. LGTM.
Tree-SHA512: 9b36b4c59bf7ad01171764ce61b1be9750fc92d105c4fe939b1a6a70027ab6300d5d2a2fc3e82f981e22c3987f2ca84e092d2e1f8463fa320af9f05048580c0a
This change also introduces an aditional buffer of 10 blocks (PRUNE_LOCK_BUFFER) that will not be pruned before the best block.
Co-authored-by: Luke Dashjr <luke-jr+git@utopios.org>
36f814c0e8 [netgroupman] Remove NetGroupManager::GetAsmap() (John Newbery)
4709fc2019 [netgroupman] Move asmap checksum calculation to NetGroupManager (John Newbery)
1b978a7e8c [netgroupman] Move GetMappedAS() and GetGroup() logic to NetGroupManager (John Newbery)
ddb4101e63 [net] Only use public CNetAddr functions and data in GetMappedAS() and GetGroup() (John Newbery)
6b2268162e [netgroupman] Add GetMappedAS() and GetGroup() (John Newbery)
19431560e3 [net] Move asmap into NetGroupManager (John Newbery)
17c24d4580 [init] Add netgroupman to node.context (John Newbery)
9b3836710b [build] Add netgroup.cpp|h (John Newbery)
Pull request description:
The asmap data is currently owned by addrman, but is used by both addrman and connman. #22791 made the data const and private (so that it can't be updated by other components), but it is still passed out of addrman as a reference to const, and used by `CNetAddress` to calculate the group and AS of the net address.
This RFC PR proposes to move all asmap data and logic into a new `NetGroupManager` component. This is initialized at startup, and the client components addrman and connman simply call `NetGroupManager::GetGroup(const CAddress&)` and `NetGroupManager::GetMappedAS(const CAddress&)` to get the net group and AS of an address.
ACKs for top commit:
mzumsande:
Code Review ACK 36f814c0e8
jnewbery:
CI failure seems spurious. I rebased onto latest master to trigger a new CI run, but whilst I was doing that, mzumsande ACKed 36f814c0e8, so I've reverted to that.
dergoegge:
Code review ACK 36f814c0e8
Tree-SHA512: 244a89cdfd720d8cce679eae5b7951e1b46b37835fccb6bdfa362856761bb110e79e263a6eeee8246140890f3bee2850e9baa7bc14a388a588e0e29b9d275175
-----
Code Reviewer Notes
Call graph of relevant functions:
UnloadBlockIndex() <-- Moved from
calls ChainstateManager::Unload()
which calls BlockManager::Unload() <-- Moved to
So calling UnloadBlockIndex() would still run this moved code. The code
will also now run when ~BlockManager gets called, which makes sense.
This is constructed before addrman and connman, and destructed afterwards.
netgroupman does not currently do anything, but will have functionality added in future commits.
Now BlockManager::LoadBlockIndex() will ACTUALLY only load BlockMan
members.
[META] In a later commit, pindexBestHeader will be moved to ChainMan as
a member
-----
Code Reviewer Notes
Call graph of relevant functions:
ChainstateManager::LoadBlockIndex() <-- Moved to
calls BlockManager::LoadBlockIndexDB()
which calls BlockManager::LoadBlockIndex() <-- Moved from
There is only one call to each of inner functions, meaning that no
behavior is changing.
fabdf9f870 Remove gui-only syscalls (MarcoFalke)
fa0c2aa826 init: Disable syscall sandbox in the bitcoin-qt process (MarcoFalke)
Pull request description:
It is basically impossible (and a bit out of scope) for us to maintain a sandbox for the qt library. I am not sure if it is possible to only sandbox a few threads in a process, but I doubt this will add no practical benefit anyway, so I am disabling the sandbox for the whole bitcoin-qt process.
See also https://github.com/bitcoin/bitcoin/pull/24690#issuecomment-1084372400
ACKs for top commit:
laanwj:
Code review ACK fabdf9f870
Tree-SHA512: 944ded03ee25f7dfd0bfeea9c3f97f575f2d470aa03b387b07f3e3bec5cb886e4aaa17e4a9fb359d3e670e6da69adc9111673d13e6561ec55b3161bb67dfe760
cccc4e879a Remove nHeightEnd and nHeight in generateBlocks helper (MarcoFalke)
fa38b1c8bd Remove buggy and confusing IncrementExtraNonce (MarcoFalke)
Pull request description:
IncrementExtraNonce has many issues:
* It is test-only code, but part of bitcoind
* It is using the block height of the tip, as opposed to the block's previous block as reference for the new height. See https://github.com/bitcoin/bitcoin/issues/24730#issuecomment-1085586193
* It has no use case in regtest testing. With a low difficulty the extra nonce won't be incremented. With a high difficulty the test-only functions are clumsy to handle anyway. For example, the generate* RPCs will return an empty array once they reached `maxtries`, as opposed to an error. Also the calls can't be aborted early unless the node shuts down completely. So I think it is fine to just remove the extra nonce functionality and leave it to the outside to implement, if needed. For example, a wrapper script can call the `generate*` RPCs once every second, to use the timestamp as extra nonce.
ACKs for top commit:
ajtowns:
ACK cccc4e879a
Tree-SHA512: d8a3989ad280ebd4b1b574159b3a396b8a42134347e6be3c88445162d86624d221c416456f45ae75aea62ed8c8a1a9bb3a2532924abca2ef7a879cb8e6b15654
