Add Gloas data column support #8682

eserilev · 2026-01-20T05:47:33Z

data columns: https://github.com/ethereum/consensus-specs/blob/master/specs/gloas/p2p-interface.md#modified-datacolumnsidecar

There are some TODOs related to gloas data column gossip verification that should be handled in a separate PR

michaelsproul · 2026-01-20T06:01:57Z

Thanks! 🙌

Looks like this would have been a lot of work 🙏

eserilev · 2026-01-20T06:20:40Z

beacon_node/beacon_chain/src/observed_data_sidecars.rs

-    /// Stores all received data indices for a given `(ValidatorIndex, Slot)` tuple.
-    items: HashMap<ProposalKey, HashSet<u64>>,
+    /// Stores all received data indices for a given `slot`.
+    items: HashMap<Slot, HashSet<u64>>,


Doesn't seem to be a reason why we use ProposalKey instead of Slot. post-gloas data columns no longer have a signed_beacon_block_header field which means we cant easily access the proposer_index

This is for DOS protection - if we use slot, an attacker could produce a column sidecar that passes gossip and prevent other column sidecars from getting processed.

Checking the spec, it looks like we need to use beacon_block_root and index
https://github.com/ethereum/consensus-specs/blob/master/specs/gloas/p2p-interface.md#data_column_sidecar_subnet_id

ah okay they key is now block root and slot
42c71c5

mergify · 2026-01-20T06:25:44Z

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

…oas-columns

jimmygchen · 2026-01-21T07:01:55Z

beacon_node/http_api/src/block_id.rs

+                })
                .collect::<Result<Vec<_>, _>>()?;

            reconstruct_blobs(&chain.kzg, data_columns, blob_indices, block, &chain.spec).map_err(


reconstruct_blobs() will return error in gloas, because signed_block_header and kzg_commitments_inclusion_proof no longer exists in the Gloas DataColumnSidecar - may be worth adding a TODO here so we don't miss this?

https://github.com/eserilev/lighthouse/blob/f25308822943798a42fff3af8c9a38d04375db3c/beacon_node/beacon_chain/src/kzg_utils.rs#L460-L470

added d7b692a

mergify · 2026-01-21T07:15:13Z

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

@eserilev

Pulling out consensus type changes from #8677. This PR covers all type changes for spec 1.7.0-alpha.1 (except for `DataColumnSidecar` changes, which is covered in @eserilev's PR #8682) Co-Authored-By: Jimmy Chen <jchen.tc@gmail.com> Co-Authored-By: Michael Sproul <michael@sigmaprime.io>

…oas-columns

jimmygchen · 2026-01-22T12:32:18Z

consensus/types/src/data/data_column_sidecar.rs


+#[derive(Debug, PartialEq, Clone)]
+pub enum Error {
+    IncorrectStateVariant,


could just add this to DataColumnSidecarError?

jimmygchen · 2026-01-22T12:38:14Z

consensus/types/src/data/data_column_sidecar.rs

+        bytes: &[u8],
+        fork_name: ForkName,
+    ) -> Result<Self, ssz::DecodeError> {
+        if fork_name.gloas_enabled() {


would it be safer to use match here, in the unlikely case if a new variant is added we don't convert into the wrong type here.

jimmygchen · 2026-01-22T12:48:48Z

testing/ef_tests/tests/tests.rs

+        SszStaticHandler::<DataColumnSidecarFulu<MinimalEthSpec>, MinimalEthSpec>::fulu_only()
            .run();
-        SszStaticHandler::<DataColumnSidecar<MainnetEthSpec>, MainnetEthSpec>::fulu_and_later()
+        SszStaticHandler::<DataColumnSidecarFulu<MainnetEthSpec>, MainnetEthSpec>::fulu_only()


should we add Gloas too?

jimmygchen · 2026-01-22T12:52:10Z

beacon_node/beacon_chain/src/fetch_blobs/fetch_blobs_beacon_adapter.rs

    }

-    pub(crate) fn blobs_known_for_proposal(
+    pub(crate) fn blobs_known_for_observation_key(


i think we only want to use this new tuple from gloas?

jimmygchen · 2026-01-22T12:52:26Z

beacon_node/beacon_chain/src/fetch_blobs/fetch_blobs_beacon_adapter.rs

    }

-    pub(crate) fn data_column_known_for_proposal(
+    pub(crate) fn data_column_known_for_observation_key(


same as above - i think fulu rules remains unchanged

jimmygchen · 2026-01-22T12:53:35Z

beacon_node/beacon_chain/src/fetch_blobs/mod.rs

-                if let Some(observed_columns) = chain_adapter_cloned.data_column_known_for_proposal(
-                    ProposalKey::new(block.message().proposer_index(), block.slot()),
-                ) {
+                if let Some(observed_columns) = chain_adapter_cloned


we may need separate logic for fulu and gloas here

jimmygchen · 2026-01-22T13:18:20Z

beacon_node/network/src/sync/block_lookups/mod.rs

+            BlockComponent::DataColumn(column) => match column.value.as_ref() {
+                DataColumnSidecar::Fulu(column) => column.block_parent_root(),
+                // TODO(gloas) we don't have a parent root post gloas, not sure what to do here
+                DataColumnSidecar::Gloas(column) => column.beacon_block_root,


this might actually be correct for lookup but need to confirm - the parent is technically the beacon block root - if we don't know the beacon block root, we'd have to look it up.

right, yeah it might be correct. Pawan said he's looking into this as he's working on lookup sync

jimmygchen · 2026-01-22T13:24:58Z

beacon_node/beacon_chain/src/beacon_chain.rs

-                slasher.accept_block_header(data_colum.signed_block_header());
+            for data_column in &data_columns {
+                // TODO(gloas) no block header post-gloas, what should we do here
+                if let DataColumnSidecar::Fulu(c) = data_column.as_data_column() {


Looks like this can be skipped and superceded by a different gossip check in gloas:

https://github.com/ethereum/consensus-specs/blob/81458afc6aad6985c533785c8d2860d87a993241/specs/gloas/p2p-interface.md?plain=1#L385

jimmygchen · 2026-01-22T13:38:12Z

beacon_node/beacon_chain/src/observed_data_sidecars.rs

        let data_indices = self
            .items
-            .entry(ProposalKey {
+            .entry(ObservationKey {


ObservationKey may have to be an enum now to support Gloas, and we could pass ObservationKey and leave it to the caller to construct

jimmygchen

Nice work!
I haven't finished but I've added comments I have so far - we mainly need to make sure it doesn't impact / change Fulu code path before we merge this, so I think we'll need to retain the ObservedDataSidecars behaviour for Fulu.

eserilev · 2026-01-22T16:01:23Z

I know a new observation key technically affects fulu code paths but it seems like unnecessary overhead to maintain separate code paths for this section of the code. Is there a reason why block root + slot won't work the same as proposer index + slot?

eserilev · 2026-01-23T02:44:29Z

re: observation key
After talking with jimmy, we're going to keep the fulu code path the same as it currently is

eserilev requested review from jxs and michaelsproul as code owners January 20, 2026 05:47

eserilev force-pushed the gloas-columns branch from d0b22fb to abd2a8c Compare January 20, 2026 05:48

Data columns for Gloas fork

1ef6104

eserilev force-pushed the gloas-columns branch from abd2a8c to 1ef6104 Compare January 20, 2026 05:48

michaelsproul added the gloas label Jan 20, 2026

eserilev added 2 commits January 19, 2026 22:05

Fix lints

8c2b63a

Add a few more todos

3088ee0

eserilev added the ready-for-review The code is ready for review label Jan 20, 2026

eserilev commented Jan 20, 2026

View reviewed changes

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 Jan 20, 2026

eserilev added 5 commits January 19, 2026 22:58

fix ef test

cad7e89

Reintroduce observation key

42c71c5

Merge branch 'unstable' of https://github.com/sigp/lighthouse into gl…

413e88a

…oas-columns

Smol fix

08a01e0

small test fix

e92dd85

jimmygchen mentioned this pull request Jan 21, 2026

Updated consensus types for Gloas 1.7.0-alpha.1 #8688

Merged

eserilev changed the base branch from gloas-consensus to unstable January 21, 2026 06:50

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 Jan 21, 2026

eserilev added 2 commits January 20, 2026 22:52

Fic

c815311

fix

f253088

jimmygchen reviewed Jan 21, 2026

View reviewed changes

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 Jan 21, 2026

eserilev added 4 commits January 21, 2026 07:18

Merge branch 'unstable' of https://github.com/sigp/lighthouse into gl…

edc58fe

…oas-columns

Add TODO

d7b692a

Fix

86f657d

Fix test

f0ead22

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 Jan 22, 2026

jimmygchen reviewed Jan 22, 2026

View reviewed changes

jimmygchen requested changes Jan 22, 2026

View reviewed changes

jimmygchen 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 Jan 22, 2026

Revert ObservationKey logic for fulu and some tests fixes\

86b105a

eserilev force-pushed the gloas-columns branch from 333563c to 86b105a Compare January 23, 2026 05:43

Add Gloas data column support #8682

Are you sure you want to change the base?

Add Gloas data column support #8682

Conversation

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

Uh oh!

Uh oh!

michaelsproul commented Jan 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!

mergify bot commented Jan 20, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mergify bot commented Jan 21, 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!

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!

jimmygchen left a comment

Choose a reason for hiding this comment

Uh oh!

eserilev commented Jan 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eserilev commented Jan 23, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

eserilev commented Jan 20, 2026 •

edited

Loading

eserilev commented Jan 22, 2026 •

edited

Loading