* focus on performance improvement in many aspects.
1. Do BlockBody verification concurrently;
2. Do calculation of intermediate root concurrently;
3. Preload accounts before processing blocks;
4. Make the snapshot layers configurable.
5. Reuse some object to reduce GC.
add
* rlp: improve decoder stream implementation (#22858)
This commit makes various cleanup changes to rlp.Stream.
* rlp: shrink Stream struct
This removes a lot of unused padding space in Stream by reordering the
fields. The size of Stream changes from 120 bytes to 88 bytes. Stream
instances are internally cached and reused using sync.Pool, so this does
not improve performance.
* rlp: simplify list stack
The list stack kept track of the size of the current list context as
well as the current offset into it. The size had to be stored in the
stack in order to subtract it from the remaining bytes of any enclosing
list in ListEnd. It seems that this can be implemented in a simpler
way: just subtract the size from the enclosing list context in List instead.
* rlp: use atomic.Value for type cache (#22902)
All encoding/decoding operations read the type cache to find the
writer/decoder function responsible for a type. When analyzing CPU
profiles of geth during sync, I found that the use of sync.RWMutex in
cache lookups appears in the profiles. It seems we are running into
CPU cache contention problems when package rlp is heavily used
on all CPU cores during sync.
This change makes it use atomic.Value + a writer lock instead of
sync.RWMutex. In the common case where the typeinfo entry is present in
the cache, we simply fetch the map and lookup the type.
* rlp: optimize byte array handling (#22924)
This change improves the performance of encoding/decoding [N]byte.
name old time/op new time/op delta
DecodeByteArrayStruct-8 336ns ± 0% 246ns ± 0% -26.98% (p=0.000 n=9+10)
EncodeByteArrayStruct-8 225ns ± 1% 148ns ± 1% -34.12% (p=0.000 n=10+10)
name old alloc/op new alloc/op delta
DecodeByteArrayStruct-8 120B ± 0% 48B ± 0% -60.00% (p=0.000 n=10+10)
EncodeByteArrayStruct-8 0.00B 0.00B ~ (all equal)
* rlp: optimize big.Int decoding for size <= 32 bytes (#22927)
This change grows the static integer buffer in Stream to 32 bytes,
making it possible to decode 256bit integers without allocating a
temporary buffer.
In the recent commit 088da24, Stream struct size decreased from 120
bytes down to 88 bytes. This commit grows the struct to 112 bytes again,
but the size change will not degrade performance because Stream
instances are internally cached in sync.Pool.
name old time/op new time/op delta
DecodeBigInts-8 12.2µs ± 0% 8.6µs ± 4% -29.58% (p=0.000 n=9+10)
name old speed new speed delta
DecodeBigInts-8 230MB/s ± 0% 326MB/s ± 4% +42.04% (p=0.000 n=9+10)
* eth/protocols/eth, les: avoid Raw() when decoding HashOrNumber (#22841)
Getting the raw value is not necessary to decode this type, and
decoding it directly from the stream is faster.
* fix testcase
* debug no lazy
* fix can not repair
* address comments
Co-authored-by: Felix Lange <fjl@twurst.com>
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 7cf56d6f064869cb62b1673f9ee437020c595391
, 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
