Update blockhash, blockhash_queue, status cache, downstream etc. using block_id

[AshwinSekar]

txs are currently signed with the blockhash - the last PoH tick.
Status cache uses blockhash for double spend verification.

In Alpenglow we don't have poh, replace blockhash with block id for alpenglow banks.

Update relevant blockhash and double spend code , ensure that RPCs and downstream tx signing doesn't break

[axaysagathiya]

This looks like a great issue to get started with! I’d love to work on it — could you please assign it to me?

[AshwinSekar]

Thanks for your interest, just assigned it to you! Feel free to drop any questions you have here and i'll respond when I can.

[AshwinSekar]

@axaysagathiya are you still planning on tackling this?

[AshwinSekar]

During replay we use this call chain to figure out if the transactions have already been processed on this fork

alpenglow/runtime/src/bank/check_transactions.rs

Lines 280 to 284 in 2a55cd1

	fn is_transaction_already_processed(
	&self,
	sanitized_tx: &impl TransactionWithMeta,
	status_cache: &BankStatusCache,
	) -> bool {

The status cache stores this info

alpenglow/runtime/src/status_cache.rs

Line 23 in 2a55cd1

pub type Status<T> = Arc<Mutex<HashMap<Hash, (usize, Vec<(KeySlice, T)>)>>>;

Bank also maintains a recent blockhash queue which is stored on chain in the sysvar account

alpenglow/runtime/src/bank.rs

Line 2406 in 2a55cd1

pub fn update_recent_blockhashes(&self) {

This is written to when we freeze a bank on this fork

alpenglow/runtime/src/bank.rs

Lines 2823 to 2825 in 2a55cd1

	fn register_recent_blockhash(&self, blockhash: &Hash, scheduler: &InstalledSchedulerRwLock) {
	// This is needed because recent_blockhash updates necessitate synchronizations for
	// consistent tx check_age handling.

We'll want to update this to use block id for alpenglow banks

This is also exposed via the RPC

alpenglow/rpc/src/rpc.rs

Line 2366 in 2a55cd1

fn get_latest_blockhash(&self, config: RpcContextConfig) -> Result<RpcResponse<RpcBlockhash>> {

Off the top of my head these are the things that need to be updated, ideally it's as simple as writing block id to the block hash queue / status cache and things "just work" but there might be unknowns.

To verify that this works we can run the invalidator cluster with bench tps and ensure that no duplicate transactions are landing.