`EmplaceCoinInternalDANGER()` incremented `cachedCoinsUsage` even when `try_emplace` did not insert (duplicate key), inflating the counter.
This is mostly reachable in tests today since `AssumeUTXO` does not overwrite.
Increment only on successful insert, and capture `coin.DynamicMemoryUsage()` before the move so accounting uses the correct value.
Fuzz: add an `EmplaceCoinInternalDANGER` path to exercise insert-only accounting.
Unit test: emplace two different coins at the same outpoint (with different `DynamicMemoryUsage()`), verify `SelfTest()` passes and `AccessCoin(outpoint)` returns the first coin.
Co-authored-by: Andrew Toth <andrewstoth@gmail.com>
Co-authored-by: w0xlt <woltx@protonmail.com>
Move the `cachedCoinsUsage` subtract in `AddCoin()` to after the `possible_overwrite` check.
Previously a throw before assignment decremented the counter without changing the entry, which corrupted accounting and later underflowed.
In `Flush()`, reset `cachedCoinsUsage` to `0` only when `BatchWrite()` succeeds and `cacheCoins` is actually cleared. In production `BatchWrite()` returns `true`, so this mostly affects tests. On failure, leave the counter unchanged to keep it in sync with the cache.
The existing `Flush()` workaround in fuzzing was also removed now that the source of the problem was fixed, so the fuzzer no longer needs `coins_view_cache.Flush()` to realign `cachedCoinsUsage` after an exception.
Replace the prior `expected_code_path` tracking with direct assertions. The role of the variable was to verify that code execution follows only expected paths, either successful addition, or if it's an exception, the message is verified and checked that overwrite was disallowed.
With these changes the counter stays consistent across success and exception paths, so we can finally remove the `UBSan` suppressions for `CCoinsViewCache` that were masking the issue.
Included a unit test as well, attempting to add a different coin to the same outpoint without allowing overwrites and make sure it throws.
We use `SelfTest()` to validates accounting, and check that the cache remains usable.
Co-authored-by: Ryan Ofsky <ryan@ofsky.org>
Co-authored-by: w0xlt <woltx@protonmail.com>
When a coin is spent via `SpendCoin()`, `cachedCoinsUsage` is already decremented and the coin's `scriptPubKey` is cleared, so `DynamicMemoryUsage()` is `0`.
`CoinsViewCacheCursor::NextAndMaybeErase()` was subtracting usage again when erasing spent entries.
Replace it with an assert that documents spent coins have zero dynamic memory usage by the time the cursor encounters them.
Remove the now-unnecessary `usage` reference from the cursor's constructor and member variables.
Updates the cmake logic to generate a separate test for each
BOOST_FIXTURE_TEST_SUITE declaration in a file, and splits coins_tests.cpp
into three separate suites so that they can be run in parallel. Also
updates the convention enforced by test/lint/lint-tests.py.
The `ccoins_add` and `ccoins_write` tests check the actual exception error messages now instead of just that they fail for the given parameters.
This enables us testing different exceptions in a more fine-grained way in later changes.
We don't need so much access to the internals of CCoinsCacheEntry, since many tests are just exercising invalid combinations this way.
This implies that `AddFlags` has private access now.
CCoinsCacheEntry provided general access to its internal flags state, even though in reality it could only be clean, fresh, dirty or fresh|dirty.
After it got dirtied we couldn't set the state back to clean by AddFlags(0) - tests were explicitly checking against that.
This commit cleans up the public interface to make this distinction cleaner and invalid behavior impossible instead of just checked by tests.
This includes the removal of redundant `inline` qualifiers (we're inside a struct).
Also renamed `self` to `pair` to simplify the upcoming commits.
Also modernized `EmplaceCoinInternalDANGER` since it was already modified.
Co-authored-by: Andrew Toth <andrewstoth@gmail.com>
Ideally all call sites should accept std::byte instead of uint8_t but those transformations are left to future PRs.
-BEGIN VERIFY SCRIPT-
sed -i --regexp-extended 's/\bParseHex\(("[^"]*")\)/\1_hex_u8/g' $(git grep -l ParseHex -- :src ':(exclude)src/test/util_tests.cpp')
sed -i --regexp-extended 's/\bParseHex<std::byte>\(("[^"]*")\)/\1_hex/g' $(git grep -l ParseHex -- :src ':(exclude)src/test/util_tests.cpp')
sed -i --regexp-extended 's/\bScriptFromHex\(("[^"]*")\)/ToScript(\1_hex)/g' src/test/script_tests.cpp
-END VERIFY SCRIPT-
Co-Authored-By: MarcoFalke <*~=`'#}+{/-|&$^_@721217.xyz>
Co-Authored-By: Ryan Ofsky <ryan@ofsky.org>
The following scripted-diff commit will replace ParseHex("...") with "..."_hex_u8, but this replacement will not work in cases where vectors are needed instead of arrays, and is not ideal in cases where std::byte is accepted.
For example, it is currently necessary to use _hex_v_u8 when calling CScript operator<< because that operator does not currently support std::array or std::byte.
Conversely, it is incorrect to use _hex_v instead of _hex in net_processing.cpp for the MakeAndPushMessage argument, because if the argument is a std::vector it is considered variable-length and serialized with a size prefix, but if the argument is a std::array or Span is it considered fixed length and serialized without a prefix.
By the same logic, it is also safe to change the NUMS_H constant in pubkey.cpp from a std::vector to std::array because it is never serialized.
Accepting any Rng in RandMoney makes tests more flexible to use a
different Rng. Also, passing in the Rng clarifies the call sites, so
that they all use g_rand_ctx explicitly and consistently.
The existing code provides two randomness mechanisms for test purposes:
- g_insecure_rand_ctx (with its wrappers InsecureRand*), which during tests is
initialized using either zeros (SeedRand::ZEROS), or using environment-provided
randomness (SeedRand::SEED).
- g_mock_deterministic_tests, which controls some (but not all) of the normal
randomness output if set, but then makes it extremely predictable (identical
output repeatedly).
Replace this with a single mechanism, which retains the SeedRand modes to control
all randomness. There is a new internal deterministic PRNG inside the random
module, which is used in GetRandBytes() when in test mode, and which is also used
to initialize g_insecure_rand_ctx. This means that during tests, all random numbers
are made deterministic. There is one exception, GetStrongRandBytes(), which even
in test mode still uses the normal PRNG state.
This probably opens the door to removing a lot of the ad-hoc "deterministic" mode
functions littered through the codebase (by simply running relevant tests in
SeedRand::ZEROS mode), but this isn't done yet.
CTxDestination is really our internal representation of an address and
doesn't really have anything to do with standard script types, so move
them to their own file.
In my benchmarks, using this pool allocator for CCoinsMap gives about
20% faster `-reindex-chainstate` with -dbcache=5000 with practically the
same memory usage. The change in max RSS changed was 0.3%.
The `validation_flush_tests` tests need to be updated because
memory allocation is now done in large pools instead of one node at a
time, so the limits need to be updated accordingly.
4275195606 De-duplicate add_coin methods to a test util helper (Jon Atack)
9d92c3d7f4 Create InsecureRandMoneyAmount() test util helper (Jon Atack)
81f5ade2a3 Move random test util code from setup_common to random (Jon Atack)
Pull request description:
- Move random test utilities from `setup_common` to a new `random` file, as many tests don't use this code.
- Create a helper to generate semi-random CAmounts up to `MONEY_RANGE` rather than only uint32, and use the helper in the unit tests.
- De-duplicate a shared `add_coin` method by extracting it to a `coins` test utility.
ACKs for top commit:
pinheadmz:
ACK 4275195606
achow101:
ACK 4275195606
john-moffett:
ACK 4275195606
Tree-SHA512: 3ed974251149c7417f935ef2f8865aa0dcc33b281b47522b0f96f1979dff94bb8527957f098fe4d210f40d715c00f29512f2ffe189097102229023b7284a3a27
Add CoinsViewOptions struct to remove ArgsManager uses from txdb.
To reduce size of this commit, this moves references to gArgs variable out of
txdb.cpp to calling code in validation.cpp. But these moves are temporary. The
gArgs references in validation.cpp are moved out to calling code in init.cpp in
later commits.
This commit does not change behavior.
1d7935b45a test: add test for coins view flush behavior using Sync() (James O'Beirne)
2c3cbd6c00 test: add use of Sync() to coins tests (James O'Beirne)
6d8affca96 test: refactor: clarify the coins simulation (James O'Beirne)
79cedc36af coins: add Sync() method to allow flush without cacheCoins drop (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 certain circumstances, we may want to flush chainstate data to disk without
emptying `cacheCoins`, which affects performance. UTXO snapshot
activation is one such case, as we populate `cacheCoins` with the snapshot
contents and want to persist immediately afterwards but also enter IBD.
See also #15265, which makes the case that under normal operation a
flush-without-erase doesn't necessarily add much benefit. I open this PR
even in light of the previous discussion because (i) flush-without-erase
almost certainly provides benefit in the case of snapshot activation (especially
on spinning disk hardware) and (ii) this diff is fairly small and gives us convenient
options for more granular cache management without changing existing policy.
See also #15218.
ACKs for top commit:
sipa:
ACK 1d7935b45a
achow101:
ACK 1d7935b45a
Sjors:
tACK 1d7935b45a
Tree-SHA512: 897583963e98661767d2d09c9a22f6019da24125558cd88770bfe2d017d924f23a9075b729e4b1febdec5b0709a38e8fa1ef94d62aa88650556b06cb4826c845
In certain circumstances, we may want to flush to disk without
emptying `cacheCoins`, which affects performance. UTXO snapshot
activation is one such case.
This method is currently unused and this commit does not
change any behavior.
Incorporates feedback from John Newbery.
Co-authored-by: Suhas Daftuar <sdaftuar@gmail.com>
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
