Peer stats were never pruned, so the peers map grew with every peer
ever seen. The EMA decay loop and stats copy iterated all historical
peers on every block/query.
Add NotifyPeerDrop(peer) that deletes the peer's stats entry. Called
from handler.unregisterPeer alongside txFetcher.Drop.
handleChainHead fetched the block by number only. If the tracker
goroutine lagged and that height was reorged before processing,
the EMA was computed from the wrong canonical block.
Use GetBlock(hash, number) with the header hash from the event to
fetch the exact block the event refers to, not whatever is currently
canonical at that height.
NotifyReceived was called before pool validation, allowing a peer
to claim deliverer credit by replaying already-included txs or
sending invalid packets.
Rename to NotifyAccepted (takes hashes, not full txs). Call it from
a new enqueueAndTrack helper in handler_eth.go that runs after
Enqueue and checks pool.Has to identify accepted txs. Only accepted
txs are credited to the delivering peer.
lastFinalNum started at 0, so the first checkFinalization after
startup iterated from block 1 to the current finalized head (~20M
blocks on mainnet) under the mutex, stalling the tracker and
potentially awarding bogus credit for ancient txs whose hashes
happened to match recently-received ones.
Seed lastFinalNum from chain.CurrentFinalBlock() in Start() so only
blocks finalized after startup are processed.
txtracker tests (7 tests):
- NotifyReceived: stats empty before chain events
- InclusionEMA: EMA increases on inclusion, decays on empty blocks
- Finalization: Finalized counter credited after finalization
- MultiplePeers: each peer credited for own txs only
- FirstDelivererWins: duplicate delivery ignored
- NoFinalizationCredit: no credit without finalization
- EMADecay: EMA approaches zero after 30 empty blocks
dropper tests (6 tests):
- FilterProtectedNoStats: nil stats → all droppable
- FilterProtectedEmptyStats: empty map → all droppable
- FilterProtectedTopPeer: top-scored peers removed from droppable
- FilterProtectedZeroScore: zero scores → no protection
- FilterProtectedOverlap: peer top in both categories → counted once
- FilterProtectedAllProtected: all droppable protected → empty list
Also fix: create peer entries during EMA update for peers with
inclusions in the current block (previously only created during
finalization, so EMA was not tracked before first finalization).
Expand the txtracker package doc to describe the tracking flow
(NotifyReceived → chain head → finalization → peer credit) and its
role as stats provider for the dropper.
Rewrite the dropper struct comment to document the full behavior
including the inclusion-based peer protection: two scoring categories
(total finalized + recent EMA), top 10% per pool, union of protected
sets.
Change the long-term protection category from total inclusions to
total finalized inclusions. Finalized txs are harder to game (require
actual block finality, not just inclusion) and represent confirmed
on-chain value.
The recent-inclusion EMA stays on chain head inclusions for
responsiveness — a peer delivering txs that appear in the latest
blocks gets quick protection without waiting for finalization.
The tracker now checks CurrentFinalBlock() on each chain head event
and credits delivering peers for all newly finalized blocks since
the last check.
Minimal txtracker that records which peer delivered each transaction
and credits peers when their transactions appear on chain. Provides
the PeerInclusionStats needed by the dropper's protection logic.
Design:
- NotifyReceived(peer, txs): records deliverer per tx hash (called
from handler_eth.go when tx bodies arrive via P2P)
- Subscribes to ChainHeadEvent, fetches block txs, credits the
delivering peer for each included tx
- Per-peer EMA of recent inclusions (alpha=0.05), updated every block
- LRU eviction at 262K entries to bound memory
- Mutex-based (not channel-based) for simplicity — the hot path
(NotifyReceived) is a fast map insert
Wired into the dropper via an adapter callback in backend.go that
converts txtracker.PeerStats to the dropper's PeerInclusionStats.
The dropper periodically disconnects random peers to create churn.
This was blind to peer quality. Add inclusion-based peer protection
using two categories:
1. Total inclusions: protects peers with the highest cumulative
count of delivered txs that were included on chain
2. Recent inclusions (EMA): protects peers with the best recent
inclusion rate, giving newly productive peers faster protection
Each category independently protects the top 10% of inbound and
top 10% of dialed peers. The union of both sets is protected. Only
peers with positive scores qualify.
The dropper defines its own PeerInclusionStats struct and callback
type (getPeerInclusionStatsFunc) so any stats provider (e.g. a
transaction tracker) can plug in without a package dependency. The
callback is nil by default (protection disabled until wired).
The protectionCategories slice is designed for easy extension —
adding a new category requires only appending a struct with a name,
scoring function, and protection fraction.
This tool is designed for the offline translation of an MPT database to
a binary trie. This is to be used for users who e.g. want to prove
equivalence of a binary tree chain shadowing the MPT chain.
It adds a `bintrie` command, cleanly separating the concerns.
This PR fixes https://github.com/ethereum/go-ethereum/issues/34623 by
changing the `vm.StateDB` interface:
Instead of `EmitLogsForBurnAccounts()` emitting burn logs, `LogsForBurnAccounts()
[]*types.Log` just returns these logs which are then emitted by the caller.
This way when tracing is used, `hookedStateDB.AddLog` will be used
automatically and there is no need to duplicate either the burn log
logic or the `OnLog` tracing hook.
ProcessBeaconBlockRoot (EIP-4788) and processRequestsSystemCall
(EIP-7002/7251) do not merge the EVM access events into the state after
execution. ProcessParentBlockHash (EIP-2935) already does this correctly
at line 290-291.
Without this merge, the Verkle witness will be missing the storage
accesses from the beacon root and request system calls, leading to
incomplete witnesses and potential consensus issues when Verkle
activates.
PathDB keys diff layers by state root, not by block hash. That means a
side-chain block can legitimately collide with an existing canonical diff layer
when both blocks produce the same post-state (for example same parent,
same coinbase, no txs).
Today `layerTree.add` blindly inserts that second layer. If the root
already exists, this overwrites `tree.layers[root]` and appends the same
root to the mutation lookup again. Later account/storage lookups resolve
that root to the wrong diff layer, which can corrupt reads for descendant
canonical states.
At runtime, the corruption is silent: no error is logged and no invariant check
fires. State reads against affected descendants simply return stale data
from the wrong diff layer (for example, an account balance that reflects one
fewer block reward), which can propagate into RPC responses and block
validation.
This change makes duplicate-root inserts idempotent. A second layer with
the same state root does not add any new retrievable state to a tree that is
already keyed by root; keeping the original layer preserves the existing parent
chain and avoids polluting the lookup history with duplicate roots.
The regression test imports a canonical chain of two layers followed by
a fork layer at height 1 with the same state root but a different block hash.
Before the fix, account and storage lookups at the head resolve the fork
layer instead of the canonical one. After the fix, the duplicate insert is
skipped and lookups remain correct.
This PR refactors the encoding rules for `AccessListsPacket` in the wire
protocol. Specifically:
- The response is now encoded as a list of `rlp.RawValue`
- `rlp.EmptyString` is used as a placeholder for unavailable BAL objects
This PR adds `GenerateTrie(db, scheme, root)` to the `triedb` package,
which rebuilds all tries from flat snapshot KV data. This is needed by
snap/2 sync so it can rebuild the trie after downloading the flat state.
The shared trie generation pipeline from `pathdb/verifier.go` was moved
into `triedb/internal/conversion.go` so both `GenerateTrie` and
`VerifyState` reuse the same code.
👋
This PR makes it possible to run "Amsterdam" in statetests. I'm aware
that they'll be failing and not in consensus with other clients, yet,
but it's nice to be able to run tests and see what works and what
doesn't
Before the change:
```
$ go run ./cmd/evm statetest ./amsterdam.json
[
{
"name": "00000019-mixed-1",
"pass": false,
"fork": "Amsterdam",
"error": "unexpected error: unsupported fork \"Amsterdam\""
}
]
```
After
```
$ go run ./cmd/evm statetest ./amsterdam.json
{"stateRoot": "0x25b78260b76493a783c77c513125c8b0c5d24e058b4e87130bbe06f1d8b9419e"}
[
{
"name": "00000019-mixed-1",
"pass": false,
"stateRoot": "0x25b78260b76493a783c77c513125c8b0c5d24e058b4e87130bbe06f1d8b9419e",
"fork": "Amsterdam",
"error": "post state root mismatch: got 25b78260b76493a783c77c513125c8b0c5d24e058b4e87130bbe06f1d8b9419e, want 0000000000000000000000000000000000000000000000000000000000000000"
}
]
```
In this PR, the Database interface in `core/state` has been extended
with one more function:
```go
// Iteratee returns a state iteratee associated with the specified state root,
// through which the account iterator and storage iterator can be created.
Iteratee(root common.Hash) (Iteratee, error)
```
With this additional abstraction layer, the implementation details can be hidden
behind the interface. For example, state traversal can now operate directly on
the flat state for Verkle or binary trees, which do not natively support traversal.
Moreover, state dumping will now prefer using the flat state iterator as
the primary option, offering better efficiency.
Edit: this PR also fixes a tiny issue in the state dump, marshalling the
next field in the correct way.
This PR implements the missing functionality for archive nodes by
pruning stale index data.
The current mechanism is relatively simple but sufficient for now:
it periodically iterates over index entries and deletes outdated data
on a per-block basis.
The pruning process is triggered every 90,000 new blocks (approximately
every 12 days), and the iteration typically takes ~30 minutes on a
mainnet node.
This mechanism is only applied with `gcmode=archive` enabled, having
no impact on normal full node.
This PR changes the blsync checkpoint init logic so that even if the
initialization fails with a certain server and an error log message is
printed, the server goes back to its initial state and is allowed to
retry initialization after the failure delay period. The previous logic
had an `ssDone` server state that did put the server in a permanently
unusable state once the checkpoint init failed for an apparently
permanent reason. This was not the correct behavior because different
servers behave differently in case of overload and sometimes the
response to a permanently missing item is not clearly distinguishable
from an overload response. A safer logic is to never assume anything to
be permanent and always give a chance to retry.
The failure delay formula is also fixed; now it is properly capped at
`maxFailureDelay`. The previous formula did allow the delay to grow
unlimited if a retry was attempted immediately after each delay period.
This is a breaking change in the opcode (structLog) tracer. Several fields
will have a slight formatting difference to conform to the newly established
spec at: https://github.com/ethereum/execution-apis/pull/762. The differences
include:
- `memory`: words will have the 0x prefix. Also last word of memory will be padded to 32-bytes.
- `storage`: keys and values will have the 0x prefix.
---------
Co-authored-by: Sina M <1591639+s1na@users.noreply.github.com>
Add persistent storage for Block Access Lists (BALs) in `core/rawdb/`.
This provides read/write/delete accessors for BALs in the active
key-value store.
---------
Co-authored-by: Jared Wasinger <j-wasinger@hotmail.com>
Co-authored-by: Gary Rong <garyrong0905@gmail.com>
In this PR, we add support for protocol version eth/70, defined by EIP-7975.
Overall changes:
- Each response is buffered in the peer’s receipt buffer when the
`lastBlockIncomplete` field is true.
- Continued request uses the same request id of its original
request(`RequestPartialReceipts`).
- Partial responses are verified in `validateLastBlockReceipt`.
- Even if all receipts for partial blocks of the request are collected,
those partial results are not sinked to the downloader, to avoid
complexity. This assumes that partial response and buffering occur only
in exceptional cases.
---------
Co-authored-by: Gary Rong <garyrong0905@gmail.com>
Co-authored-by: Felix Lange <fjl@twurst.com>
`pool.signer.Sender(tx)` bypasses the sender cache used by types.Sender,
which can force an extra signature recovery for every promotable tx
(promotion runs frequently). Use `types.Sender(pool.signer, tx)` here to
keep sender derivation cached and consistent.
This PR enables the block validation of keeper in the womir/openvm zkvm.
It also fixes some issues related to building the executables in CI.
Namely, it activates the build which was actually disabled, and also
resolves some resulting build conflicts by fixing the tags.
Co-authored-by: Leo <leo@powdrlabs.com>
Improve speed of import-history command by two orders of magnitude.
Rework ImportHistory to collect up to 2500 blocks per flush instead of
flushing after each block, reducing database commit overhead.
---------
Co-authored-by: Sina Mahmoodi <itz.s1na@gmail.com>
conn.read() used the actual UDP packet source address for
codec.Decode(), but conn.write() always used tc.remoteAddr. When the
remote node is reachable via multiple Docker networks, the packet source
IP differs from tc.remoteAddr, causing a session key lookup failure in
the codec.
Use tc.remoteAddr.String() consistently in conn.read() so the session
cache key matches what was used during Encode.
In `setOpenTelemetry`, all other fields (Enabled, Endpoint, AuthUser,
AuthPassword, InstanceID, Tags) are guarded by `ctx.IsSet()` checks, so
they only override the config file when explicitly set via CLI flags.
`SampleRatio` was the only field missing this guard, causing the flag
default (`1.0`) to always overwrite whatever was loaded from the config
file.
- Fix OpenTelemetry `SampleRatio` being unconditionally overwritten by
the CLI flag default value (`1.0`), even when the user did not pass
`--rpc.telemetry.sample-ratio`
- This caused config file values for `SampleRatio` to be silently
ignored
