Will maliciously constructed blocks lead to DDOS attacks?

[will-2012]

Trigger conditions:

reth/crates/engine/tree/src/tree/mod.rs

Line 1484 in 5c18df9

if let Some(event) = self.on_downloaded(blocks)? {

Exit at:

reth/crates/engine/tree/src/tree/mod.rs

Lines 433 to 436 in 5c18df9

	if let Err(fatal) = self.on_engine_message(msg) {
	error!(target: "engine::tree", %fatal, "insert block fatal error");
	return
	}

Possible logs are as follows:
2025-09-29T15:56:56.112734Z ERROR insert block fatal error fatal= ... 2025-09-29T15:56:56.117757Z ERROR Fatal error 2025-09-29T15:56:56.117780Z ERROR Fatal error in consensus engine 2025-09-29T15:56:56.117839Z ERROR shutting down due to error

[will-2012]

Any help or suggestions regarding these issues would be greatly appreciated!

[mattsse]

what's the fatal error message?

[will-2012]

We're working on the reth-bsc integration and are experiencing some validation errors in the BSC Parlia consensus protocol. I'm thinking we might need to address this issue at

reth/crates/engine/tree/src/tree/mod.rs

Line 2723 in 5c18df9

let validation_err = error.ensure_validation_error()?;

by adapting BSC-consensus validation error.
If I'm misunderstanding this, please correct me.

[mattsse]

hmm when this is reached this should be a consensus error

reth/crates/engine/tree/src/tree/error.rs

Lines 134 to 154 in 386eaa3

	pub fn ensure_validation_error(
	self,
	) -> Result<InsertBlockValidationError, InsertBlockFatalError> {
	match self {
	Self::Consensus(err) => Ok(InsertBlockValidationError::Consensus(err)),
	// other execution errors that are considered internal errors
	Self::Execution(err) => {
	match err {
	BlockExecutionError::Validation(err) => {
	Ok(InsertBlockValidationError::Validation(err))
	}
	// these are internal errors, not caused by an invalid block
	BlockExecutionError::Internal(error) => {
	Err(InsertBlockFatalError::BlockExecutionError(error))
	}
	}
	}
	Self::Provider(err) => Err(InsertBlockFatalError::Provider(err)),
	Self::Other(err) => Err(InternalBlockExecutionError::Other(err).into()),
	}
	}

perhaps the bsc validation uses a different variant?