additional costs in block capacity calc (#25059)

* Added additional costs to block capacity computation, and pushed alloc of CostModel all the way to the top of the call chain, instead of reallocing

* Fix two compiler errors

* Update block processing to propagate computed costs, rather than re-computing deeper in the call stack

* Clippy fix

* Reformatting fix after merge

* Add CostModel::sum_without_bpf

ledger/src/blockstore_processor.rs

@@ -65,7 +65,7 @@ use {
 };
 
 // it tracks the block cost available capacity - number of compute-units allowed
-// by max block cost limit
+// by max block cost limit.
 #[derive(Debug)]
 pub struct BlockCostCapacityMeter {
     pub capacity: u64,
@@ -163,6 +163,7 @@ fn execute_batch(
     replay_vote_sender: Option<&ReplayVoteSender>,
     timings: &mut ExecuteTimings,
     cost_capacity_meter: Arc<RwLock<BlockCostCapacityMeter>>,
+    tx_cost: u64,
 ) -> Result<()> {
     let record_token_balances = transaction_status_sender.is_some();
 
@@ -194,13 +195,14 @@ fn execute_batch(
         let remaining_block_cost_cap = cost_capacity_meter
             .write()
             .unwrap()
-            .accumulate(execution_cost_units);
+            .accumulate(execution_cost_units + tx_cost);
 
         debug!(
-            "bank {} executed a batch, number of transactions {}, total execute cu {}, remaining block cost cap {}",
+            "bank {} executed a batch, number of transactions {}, total execute cu {}, total additional cu {}, remaining block cost cap {}",
             bank.slot(),
             batch.sanitized_transactions().len(),
             execution_cost_units,
+            tx_cost,
             remaining_block_cost_cap,
         );
 
@@ -262,13 +264,15 @@ fn execute_batches_internal(
     replay_vote_sender: Option<&ReplayVoteSender>,
     timings: &mut ExecuteTimings,
     cost_capacity_meter: Arc<RwLock<BlockCostCapacityMeter>>,
+    tx_costs: &[u64],
 ) -> Result<()> {
     inc_new_counter_debug!("bank-par_execute_entries-count", batches.len());
     let (results, new_timings): (Vec<Result<()>>, Vec<ExecuteTimings>) =
         PAR_THREAD_POOL.install(|| {
             batches
                 .into_par_iter()
-                .map(|batch| {
+                .enumerate()
+                .map(|(index, batch)| {
                     let mut timings = ExecuteTimings::default();
                     let result = execute_batch(
                         batch,
@@ -277,6 +281,7 @@ fn execute_batches_internal(
                         replay_vote_sender,
                         &mut timings,
                         cost_capacity_meter.clone(),
+                        tx_costs[index],
                     );
                     if let Some(entry_callback) = entry_callback {
                         entry_callback(bank);
@@ -317,6 +322,7 @@ fn execute_batches(
     replay_vote_sender: Option<&ReplayVoteSender>,
     timings: &mut ExecuteTimings,
     cost_capacity_meter: Arc<RwLock<BlockCostCapacityMeter>>,
+    cost_model: &CostModel,
 ) -> Result<()> {
     let lock_results = batches
         .iter()
@@ -327,42 +333,64 @@ fn execute_batches(
         .flat_map(|batch| batch.sanitized_transactions().to_vec())
         .collect::<Vec<_>>();
 
-    let cost_model = CostModel::new();
     let mut minimal_tx_cost = u64::MAX;
     let mut total_cost: u64 = 0;
+    let mut total_cost_without_bpf: u64 = 0;
     // Allowing collect here, since it also computes the minimal tx cost, and aggregate cost.
     // These two values are later used for checking if the tx_costs vector needs to be iterated over.
+    // The collection is a pair of (full cost, cost without estimated-bpf-code-costs).
     #[allow(clippy::needless_collect)]
     let tx_costs = sanitized_txs
         .iter()
         .map(|tx| {
-            let cost = cost_model.calculate_cost(tx).sum();
+            let tx_cost = cost_model.calculate_cost(tx);
+            let cost = tx_cost.sum();
+            let cost_without_bpf = tx_cost.sum_without_bpf();
             minimal_tx_cost = std::cmp::min(minimal_tx_cost, cost);
             total_cost = total_cost.saturating_add(cost);
-            cost
+            total_cost_without_bpf = total_cost_without_bpf.saturating_add(cost_without_bpf);
+            (cost, cost_without_bpf)
         })
         .collect::<Vec<_>>();
 
     let target_batch_count = get_thread_count() as u64;
 
     let mut tx_batches: Vec<TransactionBatch> = vec![];
+    let mut tx_batch_costs: Vec<u64> = vec![];
     let rebatched_txs = if total_cost > target_batch_count.saturating_mul(minimal_tx_cost) {
         let target_batch_cost = total_cost / target_batch_count;
         let mut batch_cost: u64 = 0;
+        let mut batch_cost_without_bpf: u64 = 0;
         let mut slice_start = 0;
-        tx_costs.into_iter().enumerate().for_each(|(index, cost)| {
-            let next_index = index + 1;
-            batch_cost = batch_cost.saturating_add(cost);
-            if batch_cost >= target_batch_cost || next_index == sanitized_txs.len() {
-                let tx_batch =
-                    rebatch_transactions(&lock_results, bank, &sanitized_txs, slice_start, index);
-                slice_start = next_index;
-                tx_batches.push(tx_batch);
-                batch_cost = 0;
-            }
-        });
+        tx_costs
+            .into_iter()
+            .enumerate()
+            .for_each(|(index, cost_pair)| {
+                let next_index = index + 1;
+                batch_cost = batch_cost.saturating_add(cost_pair.0);
+                batch_cost_without_bpf = batch_cost_without_bpf.saturating_add(cost_pair.1);
+                if batch_cost >= target_batch_cost || next_index == sanitized_txs.len() {
+                    let tx_batch = rebatch_transactions(
+                        &lock_results,
+                        bank,
+                        &sanitized_txs,
+                        slice_start,
+                        index,
+                    );
+                    slice_start = next_index;
+                    tx_batches.push(tx_batch);
+                    tx_batch_costs.push(batch_cost_without_bpf);
+                    batch_cost = 0;
+                    batch_cost_without_bpf = 0;
+                }
+            });
         &tx_batches[..]
     } else {
+        match batches.len() {
+            // Ensure that the total cost attributed to this batch is essentially correct
+            0 => tx_batch_costs = Vec::new(),
+            n => tx_batch_costs = vec![total_cost_without_bpf / (n as u64); n],
+        }
         batches
     };
 
@@ -374,6 +402,7 @@ fn execute_batches(
         replay_vote_sender,
         timings,
         cost_capacity_meter,
+        &tx_batch_costs,
     )
 }
 
@@ -430,6 +459,7 @@ fn process_entries_with_callback(
     let mut batches = vec![];
     let mut tick_hashes = vec![];
     let mut rng = thread_rng();
+    let cost_model = CostModel::new();
 
     for entry in entries {
         match entry {
@@ -447,6 +477,7 @@ fn process_entries_with_callback(
                         replay_vote_sender,
                         timings,
                         cost_capacity_meter.clone(),
+                        &cost_model,
                     )?;
                     batches.clear();
                     for hash in &tick_hashes {
@@ -504,6 +535,7 @@ fn process_entries_with_callback(
                             replay_vote_sender,
                             timings,
                             cost_capacity_meter.clone(),
+                            &cost_model,
                         )?;
                         batches.clear();
                     }
@@ -519,6 +551,7 @@ fn process_entries_with_callback(
         replay_vote_sender,
         timings,
         cost_capacity_meter,
+        &cost_model,
     )?;
     for hash in tick_hashes {
         bank.register_tick(hash);

runtime/src/cost_model.rs

@@ -62,11 +62,15 @@ impl TransactionCost {
     }
 
     pub fn sum(&self) -> u64 {
+        self.sum_without_bpf()
+            .saturating_add(self.bpf_execution_cost)
+    }
+
+    pub fn sum_without_bpf(&self) -> u64 {
         self.signature_cost
             .saturating_add(self.write_lock_cost)
             .saturating_add(self.data_bytes_cost)
             .saturating_add(self.builtins_execution_cost)
-            .saturating_add(self.bpf_execution_cost)
     }
 }