Gloas alpha spec 8 by eserilev · Pull Request #9315 · sigp/lighthouse

eserilev · 2026-05-18T12:20:14Z

Issue Addressed

https://github.com/ethereum/consensus-specs/releases/tag/v1.7.0-alpha.8

mergify · 2026-05-18T16:38:59Z

Some required checks have failed. Could you please take a look @eserilev? 🙏

michaelsproul · 2026-05-20T02:52:22Z

+        // Set a non-zero gas_limit on latest_execution_payload_bid so the gas limit
+        // compatibility check doesn't reject all bids at genesis.
+        if let Ok(bid) = state.latest_execution_payload_bid_mut() {
+            bid.gas_limit = 30_000_000;
+        }


I wonder if this should be part of the interop spec for Gloas genesis. It seems quite important.

michaelsproul

Partial review, will complete shortly.

michaelsproul · 2026-05-20T05:06:11Z

                parent_beacon_block_root,
            }),
-            // V4 maps to V3 for SSE (slot_number is not part of the SSE spec)
+            // V4 maps to V3 for SSE (slot_number/target_gas_limit are not part of the SSE spec)


Should probably fix this at the spec level (cc @chong-he 🙏 )

michaelsproul

Have reviewed everything except the consensus changes now.

…ook more like the spec

… for invalid node variant

mergify · 2026-05-20T13:48:25Z

Some required checks have failed. Could you please take a look @eserilev? 🙏

…atch the spec and pass ef test

pawanjay176 · 2026-05-20T00:32:37Z


-    if proposal_epoch < current_epoch || proposal_epoch > current_epoch.saturating_add(1u64) {
+    if proposal_epoch < current_epoch
+        || proposal_epoch > current_epoch.saturating_add(spec.min_seed_lookahead)


ethereum/consensus-specs#5215 also says that

is_valid_proposal_slot(state, preferences) returns True, where
state is the checkpoint state at the epoch compute_epoch_at_slot(preferences.proposal_slot) - MIN_SEED_LOOKAHEAD and
the root preferences.dependent_root.

We use the head_state here which seems correct and given our previous discussion on not wanting to accept all preferences that force us to load old states. I think that is fine. Do you think this changes?

We also need to update BeaconState::is_valid_proposal_slot to use min_seed_lookahead right?

The PR also changes it in get_upcoming_proposal_slots but I don't see any direct usage of it in our codebase. This spec change is so annoying.

pawanjay176

I have reviewed the non fork choice changes and they look good. Mostly nits

pawanjay176 · 2026-05-20T18:21:21Z

+        execution_layer
+            .get_proposer_gas_limit(proposer_index)
+            .await
+            .unwrap_or(parent_gas_limit),


Shouldn't this be the default that we have configured in lighthouse instead of the parent gas limit?

Yeah agree, I think there's no harm in setting a higher gas limit here, because this is just the target. It also solves the problem I found which is that this isn't actually the parent_gas_limit anyway:

#9315 (comment)

pawanjay176 · 2026-05-20T19:05:28Z

+                        "No proposer gas limit configured, falling back to parent gas limit"
+                    );
+                }
+                proposer_gas_limit.or(Some(pre_payload_attributes.parent_gas_limit))


I couldn't find the relevant part of the spec where we fall back to the parent gas limit instead of the configured default in lighthouse

michaelsproul

Looks good. I think maybe we just add a TODO for the parent gas limit issue I raised, seeing as it is only used as a fallback, and even then is only used as a target gas limit (it can't break anything I don't think).

michaelsproul · 2026-05-21T01:27:49Z

+            cached_head
+                .snapshot
+                .beacon_state
+                .latest_execution_payload_bid()?
+                .gas_limit


I think this is wrong in the case where we build on the Full variant of the parent. The payload has not yet been applied to state in this case, so its gas limit won't be stored in state.latest_execution_payload_bid.

I think we would need to thread the proposer_head_payload_status through here, or maybe just the proposer head payload envelope, and then load the gas limit from that envelope.

michaelsproul · 2026-05-21T02:06:59Z

Actually, Pawan's suggestion is even cleaner. Get rid of parent_gas_limit from PrePayloadAttributes entirely and just use the process-level default gas limit as fallback.

eserilev added 11 commits May 18, 2026 10:24

Proposer preferences uses checkpoint root

3fe2448

Use min seed lookahead for proposer preferences verification

828b9c6

Add payload_due_bps and relevant check

08be7f5

count ptc votes from duplicated validators

93bcd6c

MIN_BUILDER_WITHDRAWABILITY_DELAY update mainnet values

b3ac2ff

track ptc participation

bb23439

allow proposer to re-org unavail blocks

d2f26d3

add target gas limit to paylaod attributes

45905dc

check bid gas limit consistency w/ target

53817f7

fmt

d92b95a

cleanup

d800535

eserilev requested a review from jxs as a code owner May 18, 2026 12:20

eserilev added the gloas label May 18, 2026

eserilev added 7 commits May 18, 2026 15:34

add unit tests for is_gas_limit_target_compatible

6a2dc05

builder deposit optimization

c9a60b1

FMT

49e83c3

Revert changes to ProposerPreferences

3878543

Fix tests

a693d56

Allow type com[plexity

f578674

update TODO

e7eb8cd

eserilev added the ready-for-review The code is ready for review label May 18, 2026

mergify Bot added waiting-on-author The reviewer has suggested changes and awaits thier implementation. and removed ready-for-review The code is ready for review labels May 18, 2026

eserilev added 6 commits May 18, 2026 21:53

Cleanup

bc98ad1

Fix fork choice issue

98ed9b9

fix

bcea7c9

Update

8926f61

Revert some more changes

e82ef46

Another revert

8d4f378

Re-intro set for validator pubkeys

21bd376

eserilev added ready-for-review The code is ready for review and removed waiting-on-author The reviewer has suggested changes and awaits thier implementation. labels May 19, 2026

Smol revert

3e4ab5a

michaelsproul reviewed May 20, 2026

View reviewed changes

eserilev added 8 commits May 20, 2026 15:03

test_slot_component_duration_calculations

f8fac6a

fork from env for gossip_verified_proposer_preferences.rs tests

02a62d3

Remove footgun in validate_on_payload_attestation

7578e15

Add todo about inlininig validate_on_payload_attestation to make it l…

48662dc

…ook more like the spec

Ensure payload_data_availability and payload_timeliness return errors…

7ba6f03

… for invalid node variant

add clarifying comments to should_build_on_full

e3b45c0

just unwrap

5603c79

saturating sub for calculating ptc participation - vote count

015dcc3

mergify Bot added waiting-on-author The reviewer has suggested changes and awaits thier implementation. and removed ready-for-review The code is ready for review labels May 20, 2026

eserilev added 5 commits May 20, 2026 16:53

safe math

c9d5f7b

Claude

a4dbdf7

warn if we fallback to parent gas limit

f27752f

parent gas limit should never be None

59068d7

check payload_received in should_extend_payload

e690423

eserilev requested a review from pawanjay176 May 20, 2026 14:41

eserilev added 2 commits May 20, 2026 19:20

move payload attestation current check to on_payload_attestation to m…

fce5898

…atch the spec and pass ef test

lint

cb2c113

pawanjay176 reviewed May 20, 2026

View reviewed changes

michaelsproul approved these changes May 21, 2026

View reviewed changes

Conversation

eserilev commented May 18, 2026

Issue Addressed

Uh oh!

mergify Bot commented May 18, 2026

Uh oh!

michaelsproul May 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

michaelsproul left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

michaelsproul left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mergify Bot commented May 20, 2026

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!

pawanjay176 left a comment

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!

michaelsproul left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

michaelsproul commented May 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

michaelsproul May 20, 2026 •

edited

Loading