validation: do not wipe utxo cache for stats/scans/snapshots

Since #28280, the cost of a non-wiping sync of the UTXO cache is only proportional to the number of dirty entries, rather than proportional to the size of the entire cache. Because of that, there is no reason to perform a wiping flush in case the contents of the cache is still useful.

Split the FlushStateMode::ALWAYS mode into a FORCE_SYNC (non-wiping) and a FORCE_FLUSH (wiping), and then use the former in scantxoutset, gettxoutsetinfo, snapshot creation.

Co-authored-by: l0rinc <pap.lorinc@gmail.com>
Co-authored-by: cedwies <141683552+cedwies@users.noreply.github.com>

contrib/tracing/log_utxocache_flush.py

@@ -46,6 +46,7 @@ FLUSH_MODES = [
     'IF_NEEDED',
     'PERIODIC',
     'FORCE_FLUSH',
+    'FORCE_SYNC',
 ]
 
 

doc/tracing.md

@@ -186,7 +186,7 @@ Is called *after* the in-memory UTXO cache is flushed.
 Arguments passed:
 1. Time it took to flush the cache microseconds as `int64`
 2. Flush state mode as `uint32`. It's an enumerator class with values
-   `0` (`NONE`), `1` (`IF_NEEDED`), `2` (`PERIODIC`), `3` (`FORCE_FLUSH`),
+   `0` (`NONE`), `1` (`IF_NEEDED`), `2` (`PERIODIC`), `3` (`FORCE_FLUSH`), `4` (`FORCE_SYNC`)
 3. Cache size (number of coins) before the flush as `uint64`
 4. Cache memory usage in bytes as `uint64`
 5. If pruning caused the flush as `bool`

src/rpc/blockchain.cpp

@@ -1043,7 +1043,7 @@ static RPCHelpMan gettxoutsetinfo()
     NodeContext& node = EnsureAnyNodeContext(request.context);
     ChainstateManager& chainman = EnsureChainman(node);
     Chainstate& active_chainstate = chainman.ActiveChainstate();
-    active_chainstate.ForceFlushStateToDisk();
+    active_chainstate.ForceFlushStateToDisk(/*wipe_cache=*/false);
 
     CCoinsView* coins_view;
     BlockManager* blockman;
@@ -2383,7 +2383,7 @@ static RPCHelpMan scantxoutset()
             ChainstateManager& chainman = EnsureChainman(node);
             LOCK(cs_main);
             Chainstate& active_chainstate = chainman.ActiveChainstate();
-            active_chainstate.ForceFlushStateToDisk();
+            active_chainstate.ForceFlushStateToDisk(/*wipe_cache=*/false);
             pcursor = CHECK_NONFATAL(active_chainstate.CoinsDB().Cursor());
             tip = CHECK_NONFATAL(active_chainstate.m_chain.Tip());
         }
@@ -3200,7 +3200,7 @@ PrepareUTXOSnapshot(
         //
         AssertLockHeld(::cs_main);
 
-        chainstate.ForceFlushStateToDisk();
+        chainstate.ForceFlushStateToDisk(/*wipe_cache=*/false);
 
         maybe_stats = GetUTXOStats(&chainstate.CoinsDB(), chainstate.m_blockman, CoinStatsHashType::HASH_SERIALIZED, interruption_point);
         if (!maybe_stats) {

src/test/fuzz/utxo_snapshot.cpp

@@ -57,7 +57,7 @@ void sanity_check_snapshot()
     // Connect the chain to the tmp chainman and sanity check the chainparams snapshot values.
     LOCK(cs_main);
     auto& cs{node.chainman->ActiveChainstate()};
-    cs.ForceFlushStateToDisk();
+    cs.ForceFlushStateToDisk(/*wipe_cache=*/false);
     const auto stats{*Assert(kernel::ComputeUTXOStats(kernel::CoinStatsHashType::HASH_SERIALIZED, &cs.CoinsDB(), node.chainman->m_blockman))};
     const auto cp_au_data{*Assert(node.chainman->GetParams().AssumeutxoForHeight(2 * COINBASE_MATURITY))};
     Assert(stats.nHeight == cp_au_data.height);

src/test/fuzz/utxo_total_supply.cpp

@@ -87,9 +87,9 @@ FUZZ_TARGET(utxo_total_supply)
         tx.vin.emplace_back(txo.first);
         tx.vout.emplace_back(txo.second.nValue, txo.second.scriptPubKey); // "Forward" coin with no fee
     };
-    const auto UpdateUtxoStats = [&]() {
+    const auto UpdateUtxoStats = [&](bool wipe_cache) {
         LOCK(chainman.GetMutex());
-        chainman.ActiveChainstate().ForceFlushStateToDisk();
+        chainman.ActiveChainstate().ForceFlushStateToDisk(wipe_cache);
         utxo_stats = std::move(
             *Assert(kernel::ComputeUTXOStats(kernel::CoinStatsHashType::NONE, &chainman.ActiveChainstate().CoinsDB(), chainman.m_blockman, {})));
         // Check that miner can't print more money than they are allowed to
@@ -99,7 +99,7 @@ FUZZ_TARGET(utxo_total_supply)
 
     // Update internal state to chain tip
     StoreLastTxo();
-    UpdateUtxoStats();
+    UpdateUtxoStats(/*wipe_cache=*/fuzzed_data_provider.ConsumeBool());
     assert(ActiveHeight() == 0);
     // Get at which height we duplicate the coinbase
     // Assuming that the fuzzer will mine relatively short chains (less than 200 blocks), we want the duplicate coinbase to be not too high.
@@ -124,7 +124,7 @@ FUZZ_TARGET(utxo_total_supply)
     circulation += GetBlockSubsidy(ActiveHeight(), Params().GetConsensus());
 
     assert(ActiveHeight() == 1);
-    UpdateUtxoStats();
+    UpdateUtxoStats(/*wipe_cache=*/fuzzed_data_provider.ConsumeBool());
     current_block = PrepareNextBlock();
     StoreLastTxo();
 
@@ -163,7 +163,7 @@ FUZZ_TARGET(utxo_total_supply)
                     circulation += GetBlockSubsidy(ActiveHeight(), Params().GetConsensus());
                 }
 
-                UpdateUtxoStats();
+                UpdateUtxoStats(/*wipe_cache=*/fuzzed_data_provider.ConsumeBool());
 
                 if (!was_valid) {
                     // utxo stats must not change

src/validation.cpp

@@ -2797,7 +2797,7 @@ bool Chainstate::FlushStateToDisk(
         bool fPeriodicWrite = mode == FlushStateMode::PERIODIC && nNow >= m_next_write;
         const auto empty_cache{(mode == FlushStateMode::FORCE_FLUSH) || fCacheLarge || fCacheCritical};
         // Combine all conditions that result in a write to disk.
-        bool should_write = empty_cache || fPeriodicWrite || fFlushForPrune;
+        bool should_write = (mode == FlushStateMode::FORCE_SYNC) || empty_cache || fPeriodicWrite || fFlushForPrune;
         // Write blocks, block index and best chain related state to disk.
         if (should_write) {
             LogDebug(BCLog::COINDB, "Writing chainstate to disk: flush mode=%s, prune=%d, large=%d, critical=%d, periodic=%d",
@@ -2871,10 +2871,10 @@ bool Chainstate::FlushStateToDisk(
     return true;
 }
 
-void Chainstate::ForceFlushStateToDisk()
+void Chainstate::ForceFlushStateToDisk(bool wipe_cache)
 {
     BlockValidationState state;
-    if (!this->FlushStateToDisk(state, FlushStateMode::FORCE_FLUSH)) {
+    if (!this->FlushStateToDisk(state, wipe_cache ? FlushStateMode::FORCE_FLUSH : FlushStateMode::FORCE_SYNC)) {
         LogWarning("Failed to force flush state (%s)", state.ToString());
     }
 }

src/validation.h

@@ -457,12 +457,13 @@ enum DisconnectResult
 class ConnectTrace;
 
 /** @see Chainstate::FlushStateToDisk */
-inline constexpr std::array FlushStateModeNames{"NONE", "IF_NEEDED", "PERIODIC", "FORCE_FLUSH"};
+inline constexpr std::array FlushStateModeNames{"NONE", "IF_NEEDED", "PERIODIC", "FORCE_FLUSH", "FORCE_SYNC"};
 enum class FlushStateMode: uint8_t {
     NONE,
     IF_NEEDED,
     PERIODIC,
     FORCE_FLUSH,
+    FORCE_SYNC,
 };
 
 /**
@@ -736,7 +737,7 @@ public:
         int nManualPruneHeight = 0);
 
     //! Flush all changes to disk.
-    void ForceFlushStateToDisk();
+    void ForceFlushStateToDisk(bool wipe_cache = true);
 
     //! Prune blockfiles from the disk if necessary and then flush chainstate changes
     //! if we pruned.

test/functional/interface_usdt_utxocache.py

@@ -110,6 +110,7 @@ FLUSHMODE_NAME = {
     1: "IF_NEEDED",
     2: "PERIODIC",
     3: "FORCE_FLUSH",
+    4: "FORCE_SYNC",
 }
 
 
@@ -385,7 +386,7 @@ class UTXOCacheTracepointTest(BitcoinTestFramework):
         bpf["utxocache_flush"].open_perf_buffer(handle_utxocache_flush)
 
         self.log.info("stop the node to flush the UTXO cache")
-        UTXOS_IN_CACHE = 2 # might need to be changed if the earlier tests are modified
+        UTXOS_IN_CACHE = 3 # might need to be changed if the earlier tests are modified
         # A node shutdown causes two flushes. One that flushes UTXOS_IN_CACHE
         # UTXOs and one that flushes 0 UTXOs. Normally the 0-UTXO-flush is the
         # second flush, however it can happen that the order changes.
@@ -415,7 +416,7 @@ class UTXOCacheTracepointTest(BitcoinTestFramework):
         bpf["utxocache_flush"].open_perf_buffer(handle_utxocache_flush)
 
         self.log.info("prune blockchain to trigger a flush for pruning")
-        expected_flushes.append({"mode": "NONE", "for_prune": True, "size": 0})
+        expected_flushes.append({"mode": "NONE", "for_prune": True, "size": BLOCKS_TO_MINE})
         self.nodes[0].pruneblockchain(315)
 
         bpf.perf_buffer_poll(timeout=500)