* Write state diff to CSV (#2)
* port statediff from 9b7fd9af80/statediff/statediff.go; minor fixes
* integrating state diff extracting, building, and persisting into geth processes
* work towards persisting created statediffs in ipfs; based off github.com/vulcanize/eth-block-extractor
* Add a state diff service
* Remove diff extractor from blockchain
* Update imports
* Move statediff on/off check to geth cmd config
* Update starting state diff service
* Add debugging logs for creating diff
* Add statediff extractor and builder tests and small refactoring
* Start to write statediff to a CSV
* Restructure statediff directory
* Pull CSV publishing methods into their own file
* Reformatting due to go fmt
* Add gomega to vendor dir
* Remove testing focuses
* Update statediff tests to use golang test pkg
instead of ginkgo
- builder_test
- extractor_test
- publisher_test
* Use hexutil.Encode instead of deprecated common.ToHex
* Remove OldValue from DiffBigInt and DiffUint64 fields
* Update builder test
* Remove old storage value from updated accounts
* Remove old values from created/deleted accounts
* Update publisher to account for only storing current account values
* Update service loop and fetching previous block
* Update testing
- remove statediff ginkgo test suite file
- move mocks to their own dir
* Updates per go fmt
* Updates to tests
* Pass statediff mode and path in through cli
* Return filename from publisher
* Remove some duplication in builder
* Remove code field from state diff output
this is the contract byte code, and it can still be obtained by querying
the db by the codeHash
* Consolidate acct diff structs for updated & updated/deleted accts
* Include block number in csv filename
* Clean up error logging
* Cleanup formatting, spelling, etc
* Address PR comments
* Add contract address and storage value to csv
* Refactor accumulating account row in csv publisher
* Add DiffStorage struct
* Add storage key to csv
* Address PR comments
* Fix publisher to include rows for accounts that don't have store updates
* Update builder test after merging in release/1.8
* Update test contract to include storage on contract intialization
- so that we're able to test that storage diffing works for created and
deleted accounts (not just updated accounts).
* Factor out a common trie iterator method in builder
* Apply goimports to statediff
* Apply gosimple changes to statediff
* Gracefully exit geth command(#4)
* Statediff for full node (#6)
* Open a trie from the in-memory database
* Use a node's LeafKey as an identifier instead of the address
It was proving difficult to find look the address up from a given path
with a full node (sometimes the value wouldn't exist in the disk db).
So, instead, for now we are using the node's LeafKey with is a Keccak256
hash of the address, so if we know the address we can figure out which
LeafKey it matches up to.
* Make sure that statediff has been processed before pruning
* Use blockchain stateCache.OpenTrie for storage diffs
* Clean up log lines and remove unnecessary fields from builder
* Apply go fmt changes
* Add a sleep to the blockchain test
* Address PR comments
* Address PR comments
* refactoring/reorganizing packages
* refactoring statediff builder and types and adjusted to relay proofs and paths (still need to make this optional)
* refactoring state diff service and adding api which allows for streaming state diff payloads over an rpc websocket subscription
* make proofs and paths optional + compress service loop into single for loop (may be missing something here)
* option to process intermediate nodes
* make state diff rlp serializable
* cli parameter to limit statediffing to select account addresses + test
* review fixes and fixes for issues ran into in integration
* review fixes; proper method signature for api; adjust service so that statediff processing is halted/paused until there is at least one subscriber listening for the results
* adjust buffering to improve stability; doc.go; fix notifier
err handling
* relay receipts with the rest of the data + review fixes/changes
* rpc method to get statediff at specific block; requires archival node or the block be within the pruning range
* review fixes
* fixes after rebase
* statediff verison meta
* fix linter issues
* include total difficulty to the payload
* fix state diff builder: emit actual leaf nodes instead of value nodes; diff on the leaf not on the value; emit correct path for intermediate nodes
* adjust statediff builder tests to changes and extend to test intermediate nodes; golint
* add genesis block to test; handle block 0 in StateDiffAt
* rlp files for mainnet blocks 0-3, for tests
* builder test on mainnet blocks
* common.BytesToHash(path) => crypto.Keaccak256(hash) in builder; BytesToHash produces same hash for e.g. []byte{} and []byte{\x00} - prefix \x00 steps are inconsequential to the hash result
* complete tests for early mainnet blocks
* diff type for representing deleted accounts
* fix builder so that we handle account deletions properly and properly diff storage when an account is moved to a new path; update params
* remove cli params; moving them to subscriber defined
* remove unneeded bc methods
* update service and api; statediffing params are now defined by user through api rather than by service provider by cli
* update top level tests
* add ability to watch specific storage slots (leaf keys) only
* comments; explain logic
* update mainnet blocks test
* update api_test.go
* storage leafkey filter test
* cleanup chain maker
* adjust chain maker for tests to add an empty account in block1 and switch to EIP-158 afterwards (now we just need to generate enough accounts until one causes the empty account to be touched and removed post-EIP-158 so we can simulate and test that process...); also added 2 new blocks where more contract storage is set and old slots are set to zero so they are removed so we can test that
* found an account whose creation causes the empty account to be moved to a new path; this should count as 'touching; the empty account and cause it to be removed according to eip-158... but it doesn't
* use new contract in unit tests that has self-destruct ability, so we can test eip-158 since simply moving an account to new path doesn't count as 'touchin' it
* handle storage deletions
* tests for eip-158 account removal and storage value deletions; there is one edge case left to test where we remove 1 account when only two exist such that the remaining account is moved up and replaces the root branch node
* finish testing known edge cases
* add endpoint to fetch all state and storage nodes at a given blockheight; useful for generating a recent atate cache/snapshot that we can diff forward from rather than needing to collect all diffs from genesis
* test for state trie builder
* minor changes/fixes
* update version meta
* if statediffing is on, lock tries in triedb until the statediffing service signals they are done using them
* update version meta
* fix mock blockchain; golint; bump patch
* increase maxRequestContentLength; bump patch
* log the sizes of the state objects we are sending
* CI build (#20)
* CI: run build on PR and on push to master
* CI: debug building geth
* CI: fix coping file
* CI: fix coping file v2
* CI: temporary upload file to release asset
* CI: get release upload_url by tag, upload asset to current relase
* CI: fix tag name
* fix ci build on statediff_at_anyblock-1.9.11 branch
* fix publishing assets in release
* bump version meta
* use context deadline for timeout in eth_call
* collect and emit codehash=>code mappings for state objects
* subscription endpoint for retrieving all the codehash=>code mappings that exist at provided height
* bump version meta
* Implement WriteStateDiffAt
* Writes state diffs directly to postgres
* Adds CLI flags to configure PG
* Refactors builder output with callbacks
* Copies refactored postgres handling code from ipld-eth-indexer
* rename PostgresCIDWriter.{index->upsert}*
* less ambiguous
* go.mod update
* rm unused
* cleanup
* output code & codehash iteratively
* had to rf some types for this
* prometheus metrics output
* duplicate recent eth-indexer changes
* migrations and metrics...
* [wip] prom.Init() here? another CLI flag?
* cleanup
* tidy & DRY
* statediff WriteLoop service + CLI flag
* [wip] update test mocks
* todo - do something meaningful to test write loop
* logging
* use geth log
* port tests to go testing
* drop ginkgo/gomega
* fix and cleanup tests
* fail before defer statement
* delete vendor/ dir
* unused
* bump version meta
* fixes after rebase onto 1.9.23
* bump version meta
* fix API registration
* bump version meta
* use golang 1.15.5 version (#34)
* bump version meta; add 0.0.11 branch to actions
* bump version meta; update github actions workflows
* statediff: refactor metrics
* Remove redundant statediff/indexer/prom tooling and use existing
prometheus integration.
* cleanup
* "indexer" namespace for metrics
* add reporting loop for db metrics
* doc
* metrics for statediff stats
* metrics namespace/subsystem = statediff/{indexer,service}
* statediff: use a worker pool (for direct writes)
* fix test
* fix chain event subscription
* log tweaks
* func name
* unused import
* intermediate chain event channel for metrics
* cleanup
* bump version meta
* update github actions; linting
* add poststate and status to receipt ipld indexes
* bump statediff version
* stateDiffFor endpoints for fetching or writing statediff object by blockhash; bump statediff version
* fixes after rebase on to v1.10.1
* update github actions and version meta; go fmt
* add leaf key to removed 'nodes'
* include Postgres migrations and schema
* service documentation
* touching up
The PR implements the --miner.notify.full flag that enables full pending block
notifications. When this flag is used, the block notifications sent to mining
endpoints contain the complete block header JSON instead of a work package
array.
Co-authored-by: AlexSSD7 <alexandersadovskyi7@protonmail.com>
Co-authored-by: Martin Holst Swende <martin@swende.se>
This adds support for EIP-2718 typed transactions as well as EIP-2930
access list transactions (tx type 1). These EIPs are scheduled for the
Berlin fork.
There very few changes to existing APIs in core/types, and several new APIs
to deal with access list transactions. In particular, there are two new
constructor functions for transactions: types.NewTx and types.SignNewTx.
Since the canonical encoding of typed transactions is not RLP-compatible,
Transaction now has new methods for encoding and decoding: MarshalBinary
and UnmarshalBinary.
The existing EIP-155 signer does not support the new transaction types.
All code dealing with transaction signatures should be updated to use the
newer EIP-2930 signer. To make this easier for future updates, we have
added new constructor functions for types.Signer: types.LatestSigner and
types.LatestSignerForChainID.
This change also adds support for the YoloV3 testnet.
Co-authored-by: Martin Holst Swende <martin@swende.se>
Co-authored-by: Felix Lange <fjl@twurst.com>
Co-authored-by: Ryan Schneider <ryanleeschneider@gmail.com>
This moves the eth config definition into a separate package, eth/ethconfig.
Packages eth and les can now import this common package instead of
importing eth from les, reducing dependencies.
Co-authored-by: Felix Lange <fjl@twurst.com>
The PR makes use of the stacktrie, which is is more lenient on resource consumption, than the regular trie, in cases where we only need it for DeriveSha
USB enumeration still occured. Make sure it will only occur if --usb is set.
This also deprecates the 'NoUSB' config file option in favor of a new option 'USB'.
This PR removes a logic in the miner, which was originally intended to help temporary testnets based on ethash from "running off into the future". If the difficulty was low, and a few computers started mining several blocks per second, the ethash rules (which demand 1s delay between blocks) would push the blocktimes further and further away.
The solution was to make the miner sleep while this happened.
Nowadays, this problem is solved instead by PoA chains, and it's recommended to let testnets and devnets be based on clique instead. The existing logic is problematic, since it can cause stalls within the miner making it difficult for remote workers to submit work if the channel is blocked on a sleep.
Credits to Saar Tochner for reporting this via the bug bounty
In miner/worker.go, there are two goroutine using channel w.newWorkCh: newWorkerLoop() sends to this channel, and mainLoop() receives from this channel. Only the receive operation is in a select.
However, w.exitCh may be closed by another goroutine. This is fine for the receive since receive is in select, but if the send operation is blocking, then it will block forever. This commit puts the send in a select, so it won't block even if w.exitCh is closed.
Similarly, there are two goroutines using channel errc: the parent that runs the test receives from it, and the child created at line 573 sends to it. If the parent goroutine exits too early by calling t.Fatalf() at line 614, then the child goroutine will be blocked at line 574 forever. This commit adds 1 buffer to errc. Now send will not block, and receive is not influenced because receive still needs to wait for the send.
A lot of times when we hit 'core' errors, example: invalid tx, the information provided is
insufficient. We miss several pieces of information: what account has nonce too high,
and what transaction in that block was offending?
This PR adds that information, using the new type of wrapped errors.
It also adds a testcase which (partly) verifies the output from the errors.
The first commit changes all usage of direct equality-checks on core errors, into
using errors.Is. The second commit adds contextual information. This wraps most
of the core errors with more information, and also wraps it one more time in
stateprocessor, to further provide tx index and tx hash, if such a tx is encoutered in
a block. The third commit uses the chainmaker to try to generate chains with such
errors in them, thus triggering the errors and checking that the generated string meets
expectations.
* miner: exit loop when downloader Done or Failed
Following the logic of the comment at the method,
this fixes a regression introduced at 7cf56d6f06
, which would allow external parties to DoS with
blocks, preventing mining progress.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: remove ineff assign (lint)
Signed-off-by: meows <b5c6@protonmail.com>
* miner: update test re downloader events
Signed-off-by: meows <b5c6@protonmail.com>
* Revert "miner: remove ineff assign (lint)"
This reverts commit eaefcd34ab4862ebc936fb8a07578aa2744bc058.
* Revert "miner: exit loop when downloader Done or Failed"
This reverts commit 23abd34265aa246c38fc390bb72572ad6ae9fe3b.
* miner: add test showing imprecise TestMiner
Signed-off-by: meows <b5c6@protonmail.com>
* miner: fix waitForMiningState precision
This helper function would return an affirmation
on the first positive match on a desired bool.
This was imprecise; it return false positives
by not waiting initially for an 'updated' value.
This fix causes TestMiner_2 to fail, which is
expected.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: remove TestMiner_2 demonstrating broken test
This test demonstrated the imprecision of the test
helper function waitForMiningState. This function
has been fixed with 6d365c2851, and this test test
may now be removed.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: fix test regarding downloader event/mining expectations
See comment for logic.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: add test describing expectations for downloader/mining events
We expect that once the downloader emits a DoneEvent,
signaling a successful sync, that subsequent StartEvents
are not longer permitted to stop the miner.
This prevents a security vulnerability where forced syncs via
fake high blocks would stall mining operation.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: use 'canStop' state to fix downloader event handling
- Break downloader event handling into event
separating Done and Failed events. We need to
treat these cases differently since a DoneEvent
should prevent the miner from being stopped on
subsequent downloader Start events.
- Use canStop state to handle the one-off
case when a downloader first succeeds.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: improve comment wording
Signed-off-by: meows <b5c6@protonmail.com>
* miner: start mining on downloader events iff not already mining
Signed-off-by: meows <b5c6@protonmail.com>
* miner: refactor miner update logic w/r/t downloader events
This makes mining pause/start logic regarding downloader
events more explicit. Instead of eternally handling downloader
events after the first done event, the subscription is closed
when downloader events are no longer actionable.
Signed-off-by: meows <b5c6@protonmail.com>
* miner: fix handling downloader events on subcription closed
Signed-off-by: meows <b5c6@protonmail.com>
* miner: (lint:gosimple) use range over chan instead of for/select
Signed-off-by: meows <b5c6@protonmail.com>
* miner: refactor update loop to remove race condition
The go routine handling the downloader events handling
vars in parallel with the parent routine, causing a
race condition.
This change, though ugly, remove the condition while
still allowing the downloader event subscription to be
closed when the miner has no further use for it (ie DoneEvent).
* miner: alternate fix for miner-flaw
Co-authored-by: meows <b5c6@protonmail.com>
This PR changes several different things:
- Adds test cases for the miner loop
- Stops the worker if it wasn't already stopped in worker.Close()
- Uses channels instead of atomics in the miner.update() loop
Co-authored-by: Felix Lange <fjl@twurst.com>
This PR significantly changes the APIs for instantiating Ethereum nodes in
a Go program. The new APIs are not backwards-compatible, but we feel that
this is made up for by the much simpler way of registering services on
node.Node. You can find more information and rationale in the design
document: https://gist.github.com/renaynay/5bec2de19fde66f4d04c535fd24f0775.
There is also a new feature in Node's Go API: it is now possible to
register arbitrary handlers on the user-facing HTTP server. In geth, this
facility is used to enable GraphQL.
There is a single minor change relevant for geth users in this PR: The
GraphQL API is no longer available separately from the JSON-RPC HTTP
server. If you want GraphQL, you need to enable it using the
./geth --http --graphql flag combination.
The --graphql.port and --graphql.addr flags are no longer available.
* eth/downloader tests: fix spurious failing test due to race between receipts/headers
* miner tests: fix travis failure on arm64
* eth/downloader: tests - store td in ancients too
* cmd, miner: add noempty-precommit flag
* cmd, miner: get rid of external flag
* miner: change bool to atomic int
* miner: fix tiny typo
Co-authored-by: Péter Szilágyi <peterke@gmail.com>
* cmd, core, eth: init tx lookup in background
* core/rawdb: tiny log fixes to make it clearer what's happening
* core, eth: fix rebase errors
* core/rawdb: make reindexing less generic, but more optimal
* rlp: implement rlp list iterator
* core/rawdb: new implementation of tx indexing/unindex using generic tx iterator and hashing rlp-data
* core/rawdb, cmd/utils: fix review concerns
* cmd/utils: fix merge issue
* core/rawdb: add some log formatting polishes
Co-authored-by: rjl493456442 <garyrong0905@gmail.com>
Co-authored-by: Péter Szilágyi <peterke@gmail.com>
* cmd/utils: make goerli the default testnet
* cmd/geth: explicitly rename testnet to ropsten
* core: explicitly rename testnet to ropsten
* params: explicitly rename testnet to ropsten
* cmd: explicitly rename testnet to ropsten
* miner: explicitly rename testnet to ropsten
* mobile: allow for returning the goerli spec
* tests: explicitly rename testnet to ropsten
* docs: update readme to reflect changes to the default testnet
* mobile: allow for configuring goerli and rinkeby nodes
* cmd/geth: revert --testnet back to ropsten and mark as legacy
* cmd/util: mark --testnet flag as deprecated
* docs: update readme to properly reflect the 3 testnets
* cmd/utils: add an explicit deprecation warning on startup
* cmd/utils: swap goerli and ropsten in usage
* cmd/geth: swap goerli and ropsten in usage
* cmd/geth: if running a known preset, log it for convenience
* docs: improve readme on usage of ropsten's testnet datadir
* cmd/utils: check if legacy `testnet` datadir exists for ropsten
* cmd/geth: check for legacy testnet path in console command
* cmd/geth: use switch statement for complex conditions in main
* cmd/geth: move known preset log statement to the very top
* cmd/utils: create new ropsten configurations in the ropsten datadir
* cmd/utils: makedatadir should check for existing testnet dir
* cmd/geth: add legacy testnet flag to the copy db command
* cmd/geth: add legacy testnet flag to the inspect command
The leaks were mostly in unit tests, and could all be resolved by
adding suitably-sized channel buffers or by restructuring the test
to not send on a channel after an error has occurred.
There is an unavoidable goroutine leak in Console.Interactive: when
we receive a signal, the line reader cannot be unblocked and will get
stuck. This leak is now documented and I've tried to make it slightly
less bad by adding a one-element buffer to the output channels of
the line-reading loop. Should the reader eventually awake from its
blocked state (i.e. when stdin is closed), at least it won't get stuck
trying to send to the interpreter loop which has quit long ago.
Co-authored-by: Felix Lange <fjl@twurst.com>
This change:
- removes the PostChainEvents method on core.BlockChain.
- sorts 'removed log' events by block number.
- fire the NewChainHead event if we inject a canonical block into the chain
even if the entire insertion is not successful.
- guarantees correct event ordering in all cases.
* travis: Enable ARM support
* Include fixes from 20039
* Add a trace to debug the invalid lookup issue
* Try increasing the timeout to see if the arm test passes
* Investigate the resolver issue
* Increase arm64 timeout for clique test
* increase timeout in tests for arm64
* Only test the failing tests
* Review feedback: don't export epsilon
* Remove investigation tricks+include fjl's feeback
* Revert the retry ahead of using the mock resolver
* Fix rebase errors
This PR adds a new unit test in miner package which will create some blocks from miner and then import into another chain. In this way, we can ensure all blocks generated by Geth miner obey consensus rules.
* cmd, eth, miner: disable advance sealing if user require
* cmd, console, miner, les, eth: wrap the miner config
* eth: remove todo
* cmd, miner: revert noadvance flag
The reason for this is: if the transaction execution is even longer
than block time, then this kind of transactions is DoS attack.
* node: close AccountsManager in new Close method
* p2p/simulations, p2p/simulations/adapters: handle node close on shutdown
* node: move node ephemeralKeystore cleanup to stop method
* node: call Stop in Node.Close method
* cmd/geth: close node.Node created with makeFullNode in cli commands
* node: close Node instances in tests
* cmd/geth, node: minor code style fixes
* cmd, console, miner, mobile: proper node Close() termination
Until this commit, when sending an RPC request that called `NewEVM`, a blank `vm.Config`
would be taken so as to set some options, based on the default configuration. If some extra
configuration switches were passed to the blockchain, those would be ignored.
This PR adds a function to get the config from the blockchain, and this is what is now used
for RPC calls.
Some subsequent changes need to be made, see https://github.com/ethereum/go-ethereum/pull/17955#pullrequestreview-182237244
for the details of the discussion.
