core/vm: reject duplicate bridge validators at Pasteur by MatusKysel · Pull Request #3623 · bnb-chain/bsc

MatusKysel · 2026-03-26T09:30:31Z

Summary

add Pasteur-specific precompile variants for BLS signature verification (0x66) and CometBFT light block validation (0x67)
reject duplicate bridge validator identities at Pasteur for both BLS pubkey aggregation inputs and light-client validator sets
add regression coverage for Pasteur precompile selection and duplicate-validator rejection paths

Details

This change introduces Pasteur-only behavior for the bridge validator precompiles.

For BLS signature verification, the Pasteur variant rejects inputs that repeat the same validator pubkey in the aggregated signer set. Pre-Pasteur behavior is preserved.

For CometBFT light block validation, the Pasteur variant validates that validator sets do not contain duplicate identities. The validation covers validator address, consensus pubkey, BLS key, and relayer
address, and is applied when decoding both the trusted consensus state and the incoming light block validator set.

Why

Duplicate bridge validators can allow malformed validator sets or signer lists to bypass uniqueness assumptions in bridge verification logic. Pasteur now enforces uniqueness at the precompile boundary instead
of relying on callers to sanitize inputs.

NathanBSC · 2026-03-31T08:33:36Z

+
+func (t validatorDuplicateTracker) check(idx int, value []byte) error {
+	key := string(value)
+	if firstIdx, ok := t.seen[key]; ok {


are bls key and RelayerAddress allowed to stay unset? represent by zero value

Yes. Those two fields are allowed to be unset upstream. In our fixed-width decode path, "unset" is represented as zero-filled bytes, so I treat empty/nil and all-zero as the same unset state and exclude only
that case from duplicate detection.

NathanBSC · 2026-03-31T08:39:57Z

 // input:
 // | chainID   | height   | nextValidatorSetHash | [{validator pubkey, voting power, relayer address, relayer bls pubkey}] |
 // | 32 bytes  | 8 bytes  | 32 bytes             | [{32 bytes, 8 bytes, 20 bytes, 48 bytes}]                               |
-func DecodeConsensusState(input []byte) (ConsensusState, error) {


how about

func DecodeConsensusState(input []byte) (ConsensusState, error)
--> func DecodeConsensusState(input []byte, requireUniqueValidators bool) (ConsensusState, error)

delete DecodeConsensusStateWithValidation

NathanBSC · 2026-03-31T08:42:29Z

 // input:
 // consensus state length | consensus state | light block |
 // 32 bytes               |                 |             |
-func DecodeLightBlockValidationInput(input []byte) (*ConsensusState, *types.LightBlock, error) {


how about

func DecodeLightBlockValidationInput(input []byte) (*ConsensusState, *types.LightBlock, error)
--> func DecodeLightBlockValidationInput(input []byte, requireUniqueValidators bool) (*ConsensusState, *types.LightBlock, error)

delete DecodeLightBlockValidationInputWithValidation

allformless · 2026-03-31T09:18:52Z

 	return "COMET_BFT_LIGHT_BLOCK_VALIDATE_HERTZ"
 }

+type cometBFTLightBlockValidateHertzMendel struct {


how about
cometBFTLightBlockValidateHertzMendel --> cometBFTLightBlockValidateMendel?
because when Mendel enabled, Hertz must be already enalbed

allformless · 2026-03-31T10:18:00Z

delay to Pasteur HF after discussion

zlacfzy · 2026-04-15T07:28:59Z

0x66 (blsSignatureVerify) is left unchanged — it is a generic BLS primitive and does not require unique pubkeys.

NathanBSC · 2026-06-16T06:09:45Z

 	common.BytesToAddress([]byte{0x1, 0x00}): &p256Verify{eip7951: true},
 }

+var PrecompiledContractsPasteur = func() PrecompiledContracts {


list precompiles directly like others

NathanBSC · 2026-06-16T06:10:18Z

 	for k := range PrecompiledContractsPrague {
 		PrecompiledAddressesPrague = append(PrecompiledAddressesPrague, k)
 	}
+	for k := range PrecompiledContractsPasteur {


move after PrecompiledContractsOsaka

hashdit-bot · 2026-06-16T07:13:58Z

Pull Request Review

This PR introduces Pasteur-specific precompile routing by adding a new PrecompiledContractsPasteur set and selecting it when rules.IsPasteur is enabled, including a Pasteur variant for light block validation at precompile 0x67. It updates CometBFT light block decoding/validation paths to optionally enforce validator uniqueness and wires this strict mode into the Pasteur validator, while keeping legacy behavior unchanged for earlier forks. The PR also adds regression tests to verify Pasteur precompile selection and rejection of duplicate validators in consensus state/light block inputs.

Sensitive Content

No sensitive content detected.

Security Issues

No serious security issues detected.

Generated by Hashdit Bot. This tool can absolutely NOT replace manual audits.

… block at Pasteur Builds on the Pasteur precompile set introduced in bnb-chain#3623. The legacy v1 Tendermint light-client precompiles 0x64 (tmHeaderValidate) and 0x65 (iavlMerkleProofValidate) are only reachable via the deprecated BC<->BSC cross-chain stack (contracts/deprecated/), which no longer exists. Rename the existing *Nano suspend handlers to tmHeaderValidateDeprecated / iavlMerkleProofValidateDeprecated (error "deprecated") and map 0x64/0x65 to them in the Pasteur set, deprecating them from Pasteur. For the v2 cometBFT light-client (0x67), the verification work scales with the light block's validator/signature count while the gas was flat. Price the Pasteur variant's RequiredGas per input byte (CometBFTLightBlockValidatePerByteGas) so gas tracks the real CPU/memory cost (cf. the per-key pricing of 0x66 blsSignatureVerify). The pre-Pasteur Hertz variant keeps the flat gas so earlier blocks replay identically. Note: the per-byte rate is tunable; raising 0x67 gas affects the live Greenfield/op-stack relayer gas budget and should be coordinated. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

… block at Pasteur (#3726) Builds on the Pasteur precompile set introduced in #3623. The legacy v1 Tendermint light-client precompiles 0x64 (tmHeaderValidate) and 0x65 (iavlMerkleProofValidate) are only reachable via the deprecated BC<->BSC cross-chain stack (contracts/deprecated/), which no longer exists. Rename the existing *Nano suspend handlers to tmHeaderValidateDeprecated / iavlMerkleProofValidateDeprecated (error "deprecated") and map 0x64/0x65 to them in the Pasteur set, deprecating them from Pasteur. For the v2 cometBFT light-client (0x67), the verification work scales with the light block's validator/signature count while the gas was flat. Price the Pasteur variant's RequiredGas per input byte (CometBFTLightBlockValidatePerByteGas) so gas tracks the real CPU/memory cost (cf. the per-key pricing of 0x66 blsSignatureVerify). The pre-Pasteur Hertz variant keeps the flat gas so earlier blocks replay identically. Note: the per-byte rate is tunable; raising 0x67 gas affects the live Greenfield/op-stack relayer gas budget and should be coordinated. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

core/vm: reject duplicate bridge validators at Mendel

ee3f3c6

mliu233 approved these changes Mar 27, 2026

View reviewed changes

annielz previously approved these changes Mar 27, 2026

View reviewed changes

NathanBSC reviewed Mar 31, 2026

View reviewed changes

allformless reviewed Mar 31, 2026

View reviewed changes

core/vm: enforce duplicate bridge validator checks at Pasteur

bb0b2cf

MatusKysel dismissed annielz’s stale review via bb0b2cf April 1, 2026 06:51

allformless added the Pasteur label Apr 1, 2026

allformless changed the title ~~core/vm: reject duplicate bridge validators at Mendel~~ core/vm: reject duplicate bridge validators at Pasteur Apr 1, 2026

mliu233 approved these changes Apr 2, 2026

View reviewed changes

zlacfzy added 2 commits April 15, 2026 15:01

core/vm: reject duplicate bridge validators at Pasteur

fb42601

core/vm: reject duplicate bridge validators at Pasteur

62ba5ac

core/vm: reject duplicate bridge validators at Pasteur

141f190

NathanBSC reviewed Jun 16, 2026

View reviewed changes

andyzhang2023 approved these changes Jun 16, 2026

View reviewed changes

contracts.go: update precompile

fdd6502

NathanBSC approved these changes Jun 16, 2026

View reviewed changes

allformless approved these changes Jun 16, 2026

View reviewed changes

flywukong approved these changes Jun 16, 2026

View reviewed changes

allformless merged commit 02122b6 into develop Jun 17, 2026
7 of 8 checks passed

allformless added this to the v1.7.4 milestone Jun 17, 2026

allformless mentioned this pull request Jun 17, 2026

release: prepare for release v1.7.4 #3729

Merged

Uh oh!

Conversation

MatusKysel commented Mar 26, 2026 • edited by allformless Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Details

Why

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

allformless commented Mar 31, 2026

Uh oh!

zlacfzy commented Apr 15, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

hashdit-bot Bot commented Jun 16, 2026

Pull Request Review

Sensitive Content

Security Issues

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

MatusKysel commented Mar 26, 2026 •

edited by allformless

Loading