We can't support a reset of the dbwrapper object when in-memory configuration is used
because it results in the permanent loss of coins. This only affects unittest
configurations (since that's the only place we use in-memory CCoinsViewDB instances).
f19fdd47a6 test: add test for CChainState::ResizeCoinsCaches() (James O'Beirne)
8ac3ef4699 add ChainstateManager::MaybeRebalanceCaches() (James O'Beirne)
f36aaa6392 Add CChainState::ResizeCoinsCaches (James O'Beirne)
b223111da2 txdb: add CCoinsViewDB::ChangeCacheSize (James O'Beirne)
Pull request description:
This is part of the [assumeutxo project](https://github.com/bitcoin/bitcoin/projects/11):
Parent PR: #15606
Issue: #15605
Specification: https://github.com/jamesob/assumeutxo-docs/tree/master/proposal
---
In the assumeutxo implementation draft (#15056), once a UTXO snapshot is loaded, a new chainstate object is created after initialization. This means that we have to reclaim some of the cache that we've allocated to the original chainstate (per `dbcache=`) to repurpose for the snapshot chainstate.
Furthermore, it makes sense to have different cache allocations depending on which chainstate is more active. While the snapshot chainstate is working to get to the network tip (and the background validation chainstate is idle), it makes sense that the snapshot chainstate should have the majority of cache allocation. And contrariwise once the snapshot has reached network tip, most of the cache should be given to the background validation chainstate.
This set of changes (detailed in the commit messages) allows us to dynamically resize the various coins caches. None of the functionality introduced here is used at the moment, but will be in the next AU PR (which introduces `ActivateSnapshot`).
`ChainstateManager::MaybeRebalanceCaches()` defines the (somewhat normative) cache allocations between the snapshot and background validation chainstates. I'd be interested in feedback if anyone has thoughts on the proportions I've set there.
ACKs for top commit:
ajtowns:
weak utACK f19fdd47a6 -- didn't find any major problems, but not super confident that I didn't miss anything
fjahr:
Code review ACK f19fdd4
ryanofsky:
Code review ACK f19fdd47a6. Only change since last review is constructor cleanup (no change in behavior). I think the suggestions here from ajtowns and others are good, but shouldn't delay merging the PR (and hold up assumeutxo)
Tree-SHA512: fffb7847fb6993dd4a1a41cf11179b211b0b20b7eb5f7cf6266442136bfe9d43b830bbefcafd475bfd4af273f5573500594aa41fff03e0ed5c2a1e8562ff9269
We'll need this to dynamically update the cache size of the existing
CCoinsViewDB instance when we create a new one during snapshot activation.
This requires us to keep the CDBWrapper instance as a pointer instead of
a reference so that we're able to destruct it and create a new instance
when the cache size changes.
Also renames `db` to `m_db` since we're already modifying each usage.
Includes feedback from Russ Yanofsky.
f9ee0f37c2 Add comments to CustomUintFormatter (Pieter Wuille)
4eb5643e35 Convert everything except wallet/qt to new serialization (Pieter Wuille)
2b1f85e8c5 Convert blockencodings_tests to new serialization (Pieter Wuille)
73747afbbe Convert merkleblock to new serialization (Pieter Wuille)
d06fedd1bc Add SER_READ and SER_WRITE for read/write-dependent statements (Russell Yanofsky)
6f9a1e5ad0 Extend CustomUintFormatter to support enums (Russell Yanofsky)
769ee5fa00 Merge BigEndian functionality into CustomUintFormatter (Pieter Wuille)
Pull request description:
The next step of changes from #10785.
This:
* Adds support for enum serialization to `CustomUintFormatter`, used in `CAddress` for service flags.
* Merges `BigEndian` into `CustomUintFormatter`, used in `CNetAddr` for port numbers.
* Converts everything (except wallet and gui) to use the new serialization framework.
ACKs for top commit:
MarcoFalke:
re-ACK f9ee0f37c2, only change is new documentation commit for CustomUintFormatter 📂
ryanofsky:
Code review ACK f9ee0f37c2. Just new commit adding comment since last review
jonatack:
Code review re-ACK f9ee0f37c2 only change since last review is an additional commit adding Doxygen documentation for `CustomUintFormatter`.
Tree-SHA512: e7a0a36afae592d5a4ff8c81ae04d858ac409388e361f2bc197d9a78abca45134218497ab2dfd6d031e0cce0ca586cf857077b7c6ce17fccf67e2d367c1b6cd4
Added are:
* Vector(arg1,arg2,arg3,...) constructs a vector with the specified
arguments as elements. The vector's type is derived from the
arguments. If some of the arguments are rvalue references, they
will be moved into place rather than copied (which can't be achieved
using list initialization).
* Cat(vector1,vector2) returns a concatenation of the two vectors,
efficiently moving elements when relevant.
Vector generalizes (and replaces) the Singleton function in
src/descriptor.cpp, and Cat replaces the Cat function in bech32.cpp
This change encapsulates UTXO set data within CChainState instances, removing
global data `pcoinsTip` and `pcoinsviewdb`. This is necessary if we want to
maintain multiple chainstates with their own rendering of the UTXO set.
We introduce a class CoinsViews which consolidates the construction of a
CCoins* hierarchy. Construction of its various pieces (db, coinscatcher,
in-memory cache) is split up so that we avoid flushing bad state to disk if
startup is interrupted.
We also introduce `CChainState::CanFlushToDisk()` which tells us when it is
safe to flush the chainstate based on this partial construction.
This commit could be broken into smaller pieces, but it would require more
ephemeral diffs to, e.g., temporarily change CCoinsViewDB's constructor
invocations.
Other changes:
- A parameter has been added to the CCoinsViewDB constructor that allows the
name of the corresponding leveldb directory to be specified.
Thanks to Russell Yanofsky and Marco Falke for helpful feedback.
-BEGIN VERIFY SCRIPT-
sed -i 's/inline std::string _(const char\* psz)/inline bilingual_str _(const char\* psz)/' src/util/translation.h
sed -i 's/return G_TRANSLATION_FUN ? (G_TRANSLATION_FUN)(psz) : psz;/return bilingual_str{psz, G_TRANSLATION_FUN ? (G_TRANSLATION_FUN)(psz) : psz};/' src/util/translation.h
sed -i 's/\b_("\([^"]\|\\"\)*")/&.translated/g' $(git grep --files-with-matches '\b_("' src)
echo Hard cases - multiline strings.
sed -i 's/"Visit %s for further information about the software.")/&.translated/g' src/init.cpp
sed -i "s/\"Only rebuild the block database if you are sure that your computer's date and time are correct\")/&.translated/g" src/init.cpp
sed -i 's/" restore from a backup.")/&.translated/g' src/wallet/db.cpp
sed -i 's/" or address book entries might be missing or incorrect.")/&.translated/g' src/wallet/wallet.cpp
echo Special case.
sed -i 's/_(COPYRIGHT_HOLDERS)/&.translated/' src/util/system.cpp test/lint/lint-format-strings.py
-END VERIFY SCRIPT-
Most includers just wanted to react to pending shutdown.
This isolates access to `fRequestShutdown` and limits access to the shutdown
api functions, including the new `AbortShutdown` for setting it to `false`.
Note I originally called `AbortShutdown` `CancelShutdown` but that name was
already taken by winuser.h
https://travis-ci.org/bitcoin/bitcoin/jobs/386913329
This change also triggered a build error in bench. Fixing it required moving LIBBITCOIN_SERVER after LIBBITCOIN_WALLET in bench_bench_bitcoin_LDADD To make
server definitions in src/net.cpp available to wallet methods in
src/wallet/wallet.cpp. Specifically, solving:
libbitcoin_wallet.a(libbitcoin_wallet_a-wallet.o): In function `CWalletTx::RelayWalletTransaction(CConnman*)':
wallet.cpp:(.text+0x3f0e): undefined reference to `CConnman::NodeFullyConnected(CNode const*)'
collect2: error: ld returned 1 exit status
https://travis-ci.org/bitcoin/bitcoin/jobs/392133581
Need for remaining init.h includes confirmed via a thorough search with a more
specific regex:
\bInterrupt\(\)|\bShutdown\(\)|\bInitLogging\(\)|\bInitParameterInteraction\(\)|\bAppInitBasicSetup\(\)|\bAppInitParameterInteraction\(\)|\bAppInitSanityChecks\(\)|\bAppInitLockDataDirectory\(\)|\bAppInitMain\(\)|\bSetupServerArgs\(\)|\bLicenseInfo\(\)|g_wallet_init_interface|init.h
Most logs should terminated with a '\n'. Some logs
are built up over multiple calls to logPrintf(), so
do not need a newline terminater. Comment all of
these 'continued' logs as a linter hing.
a192636 -blocksdir: keep blockindex leveldb database in datadir (Jonas Schnelli)
f38e4fd QA: Add -blocksdir test (Jonas Schnelli)
386a6b6 Allow to optional specify the directory for the blocks storage (Jonas Schnelli)
Pull request description:
Since the actual block files taking up more and more space, it may be desirable to have them stored in a different location then the data directory (use case: SSD for chainstate, etc., HD for blocks).
This PR adds a `-blocksdir` option that allows one to keep the blockfiles and the blockindex external from the data directory (instead of creating symlinks).
I fist had an option to keep the blockindex within the datadir, but seems to make no sense since accessing the index will (always) lead to access (r/w) the block files.
Tree-SHA512: f8b9e1a681679eac25076dc30e45e6e12d4b2d9ac4be907cbea928a75af081dbcb0f1dd3e97169ab975f73d0bd15824c00c2a34638f3b284b39017171fce2409
Using VARINT with signed types is dangerous because negative values will appear
to serialize correctly, but then deserialize as positive values mod 128.
This commit changes the VARINT macro to trigger an error by default if called
with an signed value, and updates broken uses of VARINT to pass a special flag
that lets them keep working with no change in behavior.
Currently, the READWRITE macro cannot be passed any non-const temporaries, as
the SerReadWrite function only accepts lvalue references.
Deserializing into a temporary is very common, however. See for example
things like 's >> VARINT(n)'. The VARINT macro produces a temporary wrapper
that holds a reference to n.
Fix this by accepting non-const rvalue references instead of lvalue references.
We don't propagate the rvalue-ness down, as there are no useful optimizations
that only apply to temporaries.
Then use this new functionality to get rid of many (but not all) uses of the
'REF' macro (which casts away constness).
ee4d1493e Drop upgrade-cancel callback registration for a generic "resumeable" (Matt Corallo)
Pull request description:
Instead of passing a StartShutdown reference all the way up from
txdb, give ShowProgress a "cancelable" boolean, as StartShutdown
is pretty much always what you'll want to use to cancel. Use the
same boolean to allow cancel during initial block verification.
Tree-SHA512: 515817aaa4b9e3e856200e00be9c2d44ecfa2d4f288fe3e02116105fe85de2650c13076ee7e45396ec1ce6ab45e53b0477cddda7cfdee5b3bd0589cb81a4c346
Instead of passing a StartShutdown reference all the way up from
txdb, give ShowProgress a "resumeable" boolean, which is used to
inform the user if the action will be resumed, but cancel is always
allowed by just calling StartShutdown().
