If the hash is not noexcept, unorderd_map has to assume that it can throw an exception. Thus when rehashing care needs to be taken. libstdc++ solves this by simply caching the hash value, which increases memory of each node by 8 bytes. Adding noexcept prevents this caching. In my experiments with -reindex-chainstate -stopatheight=594000, memory usage has decreased by 9.4% while runtime has increased by 1.6% due to additional hashing. Additionally, memusage::DynamicUsage() is now more accurate and does not underestimate.
and into coins.cpp. This move is necessary so that we can later include a
CCoinsViewErrorCatcher instance under CChainState.
Co-authored-by: MarcoFalke <falke.marco@gmail.com>
5d26205272 doc: explain AcceptToMemoryPoolWorker's coins_to_uncache (James O'Beirne)
Pull request description:
I found ATMPW's `coins_to_uncache` a little hard to understand (see #15264). This adds some doc for posterity.
ACKs for commit 5d2620:
jnewbery:
ACK 5d26205272
Tree-SHA512: 088508fa78012fab8680663c4e30f5cee29768416c2ca8b8b2abc29b6ac7067c5a589674f0254474a7ccc95477889d41719760f5796792bf492f51b3dd499c6c
Currently RPC commands are formatted in a way that it's easy to read
and that test/lint/check-rpc-mappings.py can parse it.
To void breaking test/lint/check-rpc-mappings.py script by running
clang-format, RPC command definitions should be disabled for clang-format.
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).
Use C++11's better capability of expressing an interface of a non-copyable class by publicly deleting its copy ctor and assignment operator instead of just declaring them private.
176c021 [qa] Test non-atomic chainstate writes (Suhas Daftuar)
d6af06d Dont create pcoinsTip until after ReplayBlocks. (Matt Corallo)
eaca1b7 Random db flush crash simulator (Pieter Wuille)
0580ee0 Adapt memory usage estimation for flushing (Pieter Wuille)
013a56a Non-atomic flushing using the blockchain as replay journal (Pieter Wuille)
b3a279c [MOVEONLY] Move LastCommonAncestor to chain (Pieter Wuille)
Tree-SHA512: 47ccc62303f9075c44d2a914be75bd6969ff881a857a2ff1227f05ec7def6f4c71c46680c5a28cb150c814999526797dc05cf2701fde1369c06169f46eccddee
This removes the possibility for GetCoin/HaveCoin/HaveCoinInCache to return
true while the respective coin is spent. By doing it across all calls, some
extra checks can be eliminated.
coins_tests is modified to call HaveCoin sometimes before and sometimes
after AccessCoin. A further change is needed because the semantics for
GetCoin slightly changed, causing a pruned entry in the parent cache to not
be pulled into the child in FetchCoin.
9417d7a33 Be much more agressive in AccessCoin docs. (Matt Corallo)
f58349ca8 Restore some assert semantics in sigop cost calculations (Matt Corallo)
3533fb4d3 Return a bool in SpendCoin to restore pre-per-utxo assert semantics (Matt Corallo)
ec1271f2b Remove useless mapNextTx lookup in CTxMemPool::TrimToSize. (Matt Corallo)
Tree-SHA512: 158a4bce063eac93e1d50709500a10a7cb1fb3271f10ed445d701852fce713e2bf0da3456088e530ab005f194ef4a2adf0c7cb23226b160cecb37a79561f29ca
3ff1fa8 Use override keyword on CCoinsView overrides (Russell Yanofsky)
24e44c3 Don't return stale data from CCoinsViewCache::Cursor() (Russell Yanofsky)
Tree-SHA512: 08699dae0925ffb9c018f02612ac6b7eaf73ec331e2f4f934f1fe25a2ce120735fa38596926e924897c203f7470e99f0a99cf70d2ce31ff428b105e16583a861
While the current implementation is pretty free, there is a lot
of possibility for this to blow up in our face with future changes,
especially as the backing map gets tweaked.
CCoinsViewCache doesn't actually support cursor iteration returning the
current contents of the cache, so raise an error when the cursor method is
called instead of returning a cursor that iterates over stale data.
Also update the gettxoutsetinfo RPC which was relying on the old behavior to be
explicit about which view it is returning data about.
This patch makes several related changes:
* Changes the CCoinsView virtual methods (GetCoins, HaveCoins, ...)
to be COutPoint/Coin-based rather than txid/CCoins-based.
* Changes the chainstate db to a new incompatible format that is also
COutPoint/Coin based.
* Implements reconstruction code for hash_serialized_2.
* Adapts the coins_tests unit tests (thanks to Russell Yanofsky).
A side effect of the new CCoinsView model is that we can no longer
use the (unreliable) test for transaction outputs in the UTXO set
to determine whether we already have a particular transaction.
The new functions are:
* CCoinsViewCache::AddCoin: Add a single COutPoint/Coin pair.
* CCoinsViewCache::SpendCoin: Remove a single COutPoint.
* AddCoins: utility function that invokes CCoinsViewCache::AddCoin for
each output in a CTransaction.
* AccessByTxid: utility function that searches for any output with
a given txid.
* CCoinsViewCache::AccessCoin: retrieve the Coin for a COutPoint.
* CCoinsViewCache::HaveCoins: check whether a non-empty Coin exists
for a given COutPoint.
The AddCoin and SpendCoin methods will eventually replace ModifyCoins
and ModifyNewCoins, AddCoins will replace CCoins::FromTx, and the new
AccessCoins and HaveCoins functions will replace their per-txid
counterparts.
Note that AccessCoin for now returns a copy of the Coin object. In a
later commit it will be change to returning a const reference (which
keeps working in all call sites).
This makes the following changes:
* In undo data and the chainstate database, the transaction nVersion
field is removed from the data structures, always written as 0, and
ignored when reading.
* The definition of hash_serialized in gettxoutsetinfo is changed to no
longer incude the nVersion field. It is renamed to hash_serialized_2
to avoid confusion. The new definition also includes transaction
height and coinbase information, as this information was missing
before.
This depends on having a CHashVerifier-based undo data checksum
verifier.
Apart from changing the definition of serialized_hash, downgrading
after using this patch is supported, as no release ever used the value
of nVersion field in UTXO entries.
We were marking coins FRESH before being sure they were not overwriting dirty undo data. This condition was never reached in existing code because undo data was always flushed before UpdateCoins was called with new transactions, but could have been exposed in an otherwise safe refactor.
Clarify in the comments the assumptions made in ModifyNewCoins.
Add ability to undo transactions to UpdateCoins unit test.
Thanks to Russ Yanofsky for suggestion on how to make logic clearer and fixing up the ccoins_modify_new test cases.
Remove the nType and nVersion as parameters to all serialization methods
and functions. There is only one place where it's read and has an impact
(in CAddress), and even there it does not impact any of the recursively
invoked serializers.
Instead, the few places that need nType or nVersion are changed to read
it directly from the stream object, through GetType() and GetVersion()
methods which are added to all stream classes.
Given that in default GetSerializeSize implementations created by
ADD_SERIALIZE_METHODS we're already using CSizeComputer(), get rid
of the specialized GetSerializeSize methods everywhere, and just use
CSizeComputer. This removes a lot of code which isn't actually used
anywhere.
For CCompactSize and CVarInt this actually removes a more efficient
size computing algorithm, which is brought back in a later commit.
MarcoFalke