update#11
Open
philipjonsen wants to merge 1750 commits into
Open
Conversation
This PR adds three cell-level kzg functions required for the sparse blobpool (eth/72). - VerifyCells: Verifies cells corresponding to proofs. This is used to verify cells received from eth/72 peers. - ComputeCells: Computes cells from blobs. This is needed because user submissions and eth/71 transaction deliveries contain blobs, while eth/72 peers expect cells. - RecoverBlobs: Recovers blobs from partial cells. This is needed to support both eth/71 and eth/72 --------- Co-authored-by: Felix Lange <fjl@twurst.com>
When `rpc.Client.Close()` is called, the TCP connection is torn down without sending a WebSocket Close frame. The server sees `websocket: close 1006 (abnormal closure): unexpected EOF` instead of a clean 1000 (normal closure). ### Root cause `websocketCodec.close()` delegates to `jsonCodec.close()` which calls `c.conn.Close()` — gorilla/websocket's `Conn.Close` explicitly "[closes the underlying network connection without sending or waiting for a close message](https://pkg.go.dev/github.com/gorilla/websocket#Conn.Close)" (per RFC 6455). ### Fix Send a WebSocket Close control frame (opcode 0x8, status 1000) before closing the underlying connection. Uses `WriteControl` with the same `encMu` mutex pattern already used by `pingLoop` for write serialization, and reuses the existing `wsPingWriteTimeout` (5s) constant. `WriteControl` errors are safe to ignore — the connection may already be broken by the time we attempt the close frame. Fixes #30482
Co-authored-by: jwasinger <j-wasinger@hotmail.com>
scheduleFetches.func1 is the biggest allocator in the long-duration profile of node (11% of total alloc_space). Each peer-iteration pre-allocated make([]common.Hash, 0, maxTxRetrievals), even for peers that end up collecting no new hashes (all their announces were already being fetched by someone else). Defer the slice allocation to the first append. Peers that collect zero hashes now pay zero allocation, which is the common case on the timeoutTrigger path where all peers with any announces are iterated.
The testPeer request counters (nAccountRequests, nStorageRequests, nBytecodeRequests, nTrienodeRequests) were plain int fields incremented with ++. These increments happen in Request* methods that are invoked concurrently by the Syncer from multiple goroutines (assignBytecodeTasks, assignStorageTasks, etc.), causing a data race reliably detected by go test -race. Change the counters to atomic.Int64 so increments and reads are synchronized without introducing a mutex. Fixes races detected in TestMultiSyncManyUseless, TestMultiSyncManyUselessWithLowTimeout, TestMultiSyncManyUnresponsive, TestSyncWithStorageAndOneCappedPeer, TestSyncWithStorageAndCorruptPeer, and TestSyncWithStorageAndNonProvingPeer.
The stateReadList field introduced by #34776 to track the state access footprint for EIP-7928 was not propagated by StateDB.Copy. Every other per-transaction field that lives alongside it (accessList, transientStorage, journal, witness, accessEvents) is copied explicitly, so this field was simply missed. After Copy the copy's stateReadList is nil while the original keeps its entries, so the nil-safe guards on StateAccessList.AddAccount / AddState silently drop every access recorded on the copy. For any post-Amsterdam code path that copies a prepared state and keeps reading from the copy, the BAL footprint becomes incomplete. Add a Copy method on bal.StateAccessList and invoke it from StateDB.Copy, matching the pattern used for accessList and accessEvents. --------- Co-authored-by: jwasinger <j-wasinger@hotmail.com>
This PR updates the BAL structure definition to the latest the spec, - Balance has been changed from [16]byte to uint256 - Storage key and value has been changed from [32]byte to uint256 - BlockAccessList has been changed from a struct to a slice of AccountChanges - TxIndex has been changed from uint16 to uint32
`StateSetWithOrigin.decode()` was missing size computation after deserializing origin data, causing `size` to remain zero after journal reload. Added the same calculation logic used in `NewStateSetWithOrigin()`.
Co-authored-by: Guillaume Ballet <3272758+gballet@users.noreply.github.com>
…ers (#34743) Save `el.Next()` before calling `plist.Remove(el)` so iteration continues correctly. Previously the loop exited after removing the first expired matcher because `Remove` invalidates the element's links. --------- Co-authored-by: Felix Lange <fjl@twurst.com>
Here, we change the EVM stack implementation to use an 'arena', i.e. a shared allocation pool for sub-call stacks. The stack is now more GC-friendly, since it is a slice of uint256 values instead of a slice of pointers. Code that pushes an item to the stack has been changed to get() the top item, then overwrite it. The PR is a rewrite/rebase of #30362. --------- Co-authored-by: Martin Holst Swende <martin@swende.se> Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de>
EIP-7825 caps the transaction gas limit at `MaxTxGas`, but after Amsterdam/EIP-8037 the transaction gas limit can include state gas reservoir in addition to the regular gas dimension. Applying the Osaka cap to the full `tx.Gas()` rejects otherwise valid Amsterdam transactions that need more than `MaxTxGas` total gas because of state gas, while their regular gas use remains within the intended limit. This changes geth to stop applying the full transaction gas cap once Amsterdam is active: - txpool stateless validation no longer rejects `tx.Gas() > MaxTxGas` under Amsterdam - legacy pool reorg cleanup does not purge high-total-gas transactions at the Osaka transition if Amsterdam is also active - execution precheck mirrors the txpool behavior and does not reject high-total-gas messages under Amsterdam The block gas limit check remains in place, so transactions still cannot request more total gas than the current block gas limit. Validation run: ``` go test ./core/txpool ./core/txpool/legacypool go test ./core -run TestStateProcessorErrors ``` --------- Co-authored-by: Gary Rong <garyrong0905@gmail.com>
- Fixes an error shadowing issue in the deliver() function, where a stale result from GetDeliverySlot caused the original failure to be overwritten by errStaleDelivery. - Adds errInvalidBody and errInvalidReceipt to the downloader error checks to properly drop peers who sent invalid responses. --------- Co-authored-by: Felix Lange <fjl@twurst.com>
…st (#34639) The layer-5 diff condition used `i > 50 || i < 85`, which is true for almost all keys in the 0..255 loop. Use `i > 50 && i < 85` so layer 5 only covers the intended band (51..84), consistent with the snapshot iterator test fix.
Next() function in RawIterator returned true on decompression errors.Now it returns false on those cases. Redundant error check on cmd/era/main.go is also removed. --------- Co-authored-by: Bosul Mun <bsbs8645@snu.ac.kr>
This PR addresses one of the biggest performance issue with binary tries: storing each internal node individually bloats the index, the disk, and triggers a lot of write amplifications. To fix this issue, this PR serializes groups of nodes together. Because we are still looking for the ideal group size, the "depth" of the group tree is made a parameter, but that will be removed in the future, once the perfect size is known. This is a rebase of #33658 --------- Co-authored-by: Copilot <copilot@github.com>
The mux tracer fanned out every standard hook to its children but never
forwarded OnSystemCall{Start,End}. Tracers that rely on these - like
`logger.jsonLogger`, which uses the start hook to silence its opcode
hook for the duration of a system call - never got the signal when
wrapped behind a mux.
In evm t8n, combining `--trace` with `--opcode-count` (default for geth
with exec specs) produces exactly that wrapping. The first system call
(e.g. `ProcessBeaconBlockRoot`) then fires `OnOpcode` on the json logger
before any `OnTxStart` has run, dereferencing a nil env and crashing
t8n.
Forward both hooks through the mux. The V2 fan-out falls back to V1 for
children that only implement the legacy hook, mirroring the precedence
already used in `core/state_processor.go`.
This updates the typed `ethclient` model for `eth_simulateV1` call results to include `maxUsedGas`, matching the field already returned by the server-side RPC response. Follow-up to #32789.
Fixes a condition in a snapshot-related test.
Disables the recently added log indexer from a simulated backend. In most cases the log indexer is not required and unindexed search should be fast enough. Fixes #32552.
Apply block overrides to header in eth_call so EIP-1559 fee fields use the correct overridden basefee.
the gapped queue cap was effectively per-sender rather than total — a sender pool spread across enough distinct addresses could grow `p.gapped` well past `maxGapped`, defeating the resource bound. `maxGapped` was being compared against `len(p.gapped)`, which is a `map[address][]tx` and counts unique senders, not queued txs. Switched the check to `len(p.gappedSource)` (keyed by tx hash, so its length is the real total). Also wired up a `blobpool/gapped/count` gauge plus `promoted`, `evicted`, and `gappedfull` meters so queue size and churn are actually observable in prod.
Fixes the muxTracer to correctly forward events to v2 state hooks, i.e. `OnCodeChangeV2` and `OnNonceChangeV2`.
Re-exports errors in bind package. --------- Co-authored-by: Marius van der Wijden <m.vanderwijden@live.de>
Include the actual and expected request IDs in the devp2p test response failure message. This also fixes the typo in the diagnostic so mismatched responses report a useful error. --------- Co-authored-by: Bosul Mun <bsbs8645@snu.ac.kr>
SignTx populated BlobHashes and the sidecar fields (Blobs/Commitments/ Proofs) for a blob transaction but never set args.BlobFeeCap. As a result the external (clef) signer received maxFeePerBlobGas:null and signed a transaction inconsistent with the one passed in, silently dropping the blob fee cap. Set args.BlobFeeCap from tx.BlobGasFeeCap() so the signing request faithfully reflects the input transaction. This mirrors the existing handling of the other blob-tx fields.
The LookupRandom RPC handler obtains a node iterator from RandomNodes() but never closes it. RandomNodes() returns a lookupIterator backed by a cancelable context derived from the listener's lifetime context (newLookupIterator -> context.WithCancel). The iterator's Close() is what calls the cancel func, so each LookupRandom call that never closes leaks the cancel func (and any in-flight lookup goroutine) until the discv4 listener shuts down. When the listener serves the RPC API (discv4 --rpc) it is long-lived, so repeated LookupRandom calls accumulate the leak. Close the iterator when the handler returns. This matches how the crawler already releases the same RandomNodes() iterators (crawl.go closes every iterator it consumes).
EIP-7928 requires that each address appears exactly once in the block access list. In the `BlockAccessList.Validate`, the strict `isStrictlySortedFunc` is required to detect the duplicated accounts.
This PR addresses an issue in the eth71 `BlockAccessListsMsg` handler, specifically: - if the requested bal is not accessible in the server side, 0x80 (EmptyString) will be returned as the marker - at the client side, old message definition `rlp.RawList[RawBlockAccessList]` assumes all the elements are List - the message with 0x80 (kind = string) won't be decoded correctly - the peer will be disconnected The message definition has been changed to `rlp.RawList[rlp.RawValue]`, which is aligned with the one in SNAP/2 protocol.
Increment `dropCount` on every TALKREQ dropped due to overload timeout. Previously the counter was only updated when the throttled warning log was emitted.
The drop helper uses sort.Search on the sorted wallet list, but the returned index can point at the next greater wallet when the target URL is missing. Fix this by rewriting the drop to use slices.DeleteFunc and a map. --------- Co-authored-by: Felix Lange <fjl@twurst.com>
This fixes an error where a failed key decode prints its arguments in the wrong order.
`FairMix.AddSource` takes ownership of the supplied iterator, but if the mixer has already been closed it currently returns without closing that iterator. Close the rejected source in this shutdown path so callers racing with `Close` do not leak iterator resources.
This PR fixes five remaining differences between master and glamsterdam-devnet-6: - RegularCost for Auths has been increased, since 8037 increases the WARM_ACCESS cost - Floor is anchored to the transaction base cost, not the intrinsic cost - Auth destinations need to be recorded in the BAL before the call is executed - Change the place where intrinsic gas is verified in calls to delegated addresses - Refund state gas directly to reservoir in outer tx frame --------- Co-authored-by: Gary Rong <garyrong0905@gmail.com>
## Summary - Add the missing `forks.Amsterdam` case to `ChainConfig.Timestamp`. - Callers can now look up the Amsterdam fork activation time. ## Test plan - [x] `go test -short ./params/`
The HTTP JSON-RPC server caps request bodies at a hardcoded 5 MB (`defaultBodyLimit`), with no way to raise it. EEST bloatnet depth benchmarks post worst-case batch requests larger than that during fill-stateful, which the server rejects with `413 Request Entity Too Large`. This adds `node.Config.HTTPBodyLimit` (default 5 MB), wired through to the HTTP/WS server config, and exposes it via `--rpc.http-body-limit` (value in **megabytes**, default 5). Default behaviour is unchanged.
Populate the Index field on captured logs, matching callTracer behaviour.
Passing `--fuzz=false` currently still selects blocktest fuzz output mode because the command checks whether the flag was set. This switches the blocktest fuzz-mode checks to use the boolean flag value, so explicit false keeps the normal output path while `--fuzz` still enables fuzz-oriented reporting.
…35300) This PR clears `journal.writer` immediately after `Close()` in `rotate` and `setupWriter`, before checking the error. This prevents leaving a stale file handle that could be reused if close fails. --------- Co-authored-by: Bosul Mun <bsbs8645@snu.ac.kr>
This is an edge case found by @weiihann. Under 8038, the cold storage access cost is increased to 3,000 gas, which exceeds the sentry check threshold. Therefore, the sentry check no longer guarantees that the remaining gas is sufficient to cover a cold slot access. Therefore, an additional access affordability check is added to eliminate the potential DoS vector.
This PR applies the 7997 irregular state transition in t8n, block building, simulation and tracing.
) Closes #35314. ### Problem `TransactionArgs.ToTransaction` selects `SetCodeTxType` when an `authorizationList` is present, but then unconditionally downgrades to `LegacyTxType` when `gasPrice` is set: ```go case args.AuthorizationList != nil || defaultType == types.SetCodeTxType: usedType = types.SetCodeTxType ... if args.GasPrice != nil { usedType = types.LegacyTxType } ``` A legacy transaction cannot carry an EIP-7702 authorization list, so the list is silently dropped. `eth_sendTransaction`, `eth_signTransaction` and `eth_fillTransaction` can then return a plain legacy transaction/hash even though the requested delegation update or revocation was never included. The downgrade also masks a latent issue: `setFeeDefaults` returns early once `gasPrice` is set (without `authorizationList`) and never fills `MaxFeePerGas`/`MaxPriorityFeePerGas`. Building the `SetCodeTx` without the downgrade would hit `uint256.MustFromBig((*big.Int)(args.MaxFeePerGas))` with a nil fee cap and panic. Erroring early is therefore the correct behavior. ### Fix Reject the `gasPrice` + `authorizationList` combination in `setFeeDefaults`, mirroring the existing `gasPrice` / EIP-1559 guard, so the request fails explicitly instead of producing a transaction that omits the authorization list. ### Test Added a `setFeeDefaults` case asserting the new error. Verified fail-on-main: with the guard reverted the test fails (`expected error: both gasPrice and authorizationList specified`); with the guard it passes. Full `internal/ethapi` package, `go vet` and `gofmt` are clean. ### Note The same silent-drop applies to `gasPrice` + `blobVersionedHashes` (blob transactions also cannot be legacy). I scoped this PR to the reported `authorizationList` case; happy to extend the guard to blob hashes in the same spot if preferred.
This PR does a few things: - reject `debug_setHead` if the target is even before the pivot block (if non-nil) - reject `debug_setHead` if in path mode, the target is not recoverable - decouple the chain rewinding and state recovery in path mode and recover the state in one shot --------- Co-authored-by: jonny rhea <5555162+jrhea@users.noreply.github.com>
## Summary Sanity fixes surfaced by running the EELS `tests@v20.0.0` fixture release (63,109 blockchain tests) through `evm blocktest`. Also bumps CI to consume the new release: `build/checksums.txt` now points at `tests@v20.0.0` / `fixtures.tar.gz` from `ethereum/execution-specs` (supersedes the archived EEST repo's `fixtures_develop`). --------- Co-authored-by: Gary Rong <garyrong0905@gmail.com>
…ion stats (#34702) The peer dropper periodically disconnects random peers to create churn. This was previously blind to peer quality. This PR adds peer-score based peer protection, handling the multi-dimensionality problem of peer scoring through the concept of protected peer pools. --------- Signed-off-by: Csaba Kiraly <csaba.kiraly@gmail.com> Co-authored-by: healthykim <bsbs8645@snu.ac.kr>
Updates ckzg to the newest version
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.